Horticulture

A total of 6,236 SNPs were detected as being significantly associated with the host Vitis using the Bonferroni correction for multiple hypothesis testing with p-values <0.05. However, this still does not exclude the possibility of significance via phylogenetic proximity across the tree. Given the few clades of Vitis-associated bacteria, this dataset did not offer the power to use the most conservative method used by Scoary, the worst pairwise comparison p-value, which would identify only SNPs that have arisen independently across the phylogeny. However, data presented here , show genes that have a significant corrected P-value, and also have a best pairwise comparison P <=.125, showing some indication of independent emergence. 22 SNPs fit those criteria. Out of those 22, there were 9 genes in which several SNPs within them emerged as significant, seedling starter pot likely due to linkage disequilibrium. Amongst the then total of 13 genes identified with significant SNPs, 5 are only identified to the level of hypothetical protein. Most genes that had multiple SNPs detected as significant also had identical frequencies of those SNPs across the populations, showing high linkage of those sites. Due to this, only one SNP is shown per gene and included in the enrichment tests.

The 8 non-hypothetical significant genes are azu , carA, , clpP , nadD , mutY , ubiJ , exbD_1/2 , IpxD3/4 . KEGG orthologies show that 5 of the 8 identified and named genes are classified in the metabolism family in terms of their molecular functions . All identified SNPs were present at lower frequencies in strains from Vitis than other hosts, with a mean frequency of 0.047 in Vitis derived strains and a mean frequency of 0.49 in non-Vitis derived strains. Gene gain and loss was also tested using Scoary for host associations and the same pvalue corrections to go from 473 genes with Bonferroni p <0.05 to 37 genes that also had a best pairwise comparison p <= 0.125 . Out of the 37 identified genes, 33 are only identified as hypothetical proteins. The four named genes are xerC_1/2 , hcaB , mdtA_1/2/3 , and a cluster identified using Panaroo that includes the three genes CnrA , swrC_2 , and acrF_2 .These genes each belong to a different KEGG orthology .Information about X. fastidiosa ssp. fastidiosa infections in coffee in Central America is still limited – little symptomatology is associated with the disease; however, there are persistent infections in the field . The results of this work suggest that the host jump included adaptation of the X. fastidiosa strains to Vitis and lessened the ability of the U.S. strains to infect C. arabica. However, the ability for the pathogen to move throughout the C. arabica plants’ xylem vessels demonstrates a higher infectivity than expected in a fully resistant plant.

Movement of the strains away from the inoculation point was observed, which shows the ability of the bacteria to successfully degrade the pit membrane, which is often a barrier to colonization. However, in contrast to that finding, there was a reduction of positive-testing C. arabica plants over the course of the year after the inoculation. This demonstrates a lower infectivity in C. arabica than the strains have to V. vinifera where infections in a greenhouse are sustained post-inoculation. This could demonstrate a reduction in the ability to create chronic infections in C. arabica by the Vitis-adapted strains. It is possible that the chronic infections of X. fastidiosa in C. arabica plants observed in situ may be caused by high rates of re-inoculation by insect vectors rather than strain level variation in infectivity. That is not the case for ssp. fastidiosa in V. vinifera, just one inoculation is sufficient for high virulence, and only cold winter temperatures have been known to cure infections that are otherwise chronic . The virulence of the U.S. strains to C. arabica is not as high as to V. vinifera, as shown by the lack of severe symptoms in C. arabica. Given that, this study does not offer evidence that the California ssp. fastidiosa strains are generally more virulent than the ancestral strains in Central America. While not a likely scenario, there is a possibility that instead of experiencing adaptation to a specific host, the introduced strains became generally more virulent, which has been hypothesized about the globally spreading ssp. pauca strain infecting olive in Italy . In this paper, we present hypervirulence as a possible scenario, but as it is not supported by the data, we were able to rule it out. Symptom development in infected C. arabica plants consists of minor stunted growth, compared to the severe leaf scorch, matchstick petioles, shriveled fruit, and often plant death that occurs in V. vinifera. There are records of virulence of X. fastidiosa in C. arabica, such as evaluation of ssp. pauca in Brazil. In inoculation experiments with those strains, C. arabica may still develop symptoms slowly, and the proportion of positive plants does also reduce, however after 8 months the percentage positive was around 30% , not entirely dissimilar to the results from this project, which extends some uncertainty about this system.

All methods used detected genetic signatures of adaptation, and many genetic candidates were identified by multiple methods. We were able to identify genes and SNPs associated with the host Vitis as well as genes under positive selection in strains isolated from Vitis. These genes included many hypothetical proteins, however genes with known functions pertaining to infections by X. fastidiosa were also identified.The genes identified by multiple methods included those whose functions in X. fastidiosa have been previously investigated, while many still have unknown functions . ClpX, the gene for the ATP-dependent protease ATP binding subunit, was previously identified as being upregulated 4x during induction of biofilm condition , a physiological state that is vital for virulence and vector colonization . Mutations in the copper related gene copA have been found to drastically alter copper tolerance in X. fastidiosa, which is vital for agricultural survival given the frequency of copper in treating fungal infections in the vineyard, a fungicide that has been in use since the 18th century . It is conceivable that after a host jump into a vineyard, it would be necessary for pathogens to survive higher levels of copper exposure. DegP has been found to be upregulated upon heat shock in X. fastidiosa . TolB encodes for a translocation protein involved in membrane integrity and has been shown to be important for biofilm development, an important aspect of pathogenicity . These are among a suite of other genes that are both hypothetical proteins, or just understudied in X. fastidiosa but show evidence of being involved in the process of this climate and host shift. Now that these genes have been identified, they are prime candidates for targeting in future experiments to determine their effects on host range and climate adaptation. This study also includes one hypothesis that lies outside the general narrative of the introduction event, round nursery pots namely the evaluation of infectivity of ssp. multiplex towards C. arabica, which has not been tested before. While in California, ssp. multiplex has never been found infecting grapevine in the field, it has been shown to generate non persistent infections, similar to what we observed in C. arabica, in the greenhouse . Recently, infections of grapevine by ssp. fastidiosa have been detected in the field in Virginia . While the ssp. multiplex infections were not highly virulent, they were just as persistent as the ssp. fastidiosa strains. In conclusion, we have identified a suite of genes that are related to a host switch to Vitis with a corresponding reduction in the ability to infect an ancestral host. These data support the hypothesis that the shift was not a host range expansion of the subspecies, but a reduction of ability to infect a former host while optimizing the ability to infect a new host species.The results of in-situ experiments can differ greatly from those conducted in controlled laboratory conditions, especially in applied biological systems. In pathology specifically, there are often ethical and regulatory hurdles that prevent the introduction of disease-causing pathogens into naïve systems, whether it is for human health, food security, or in natural systems.

Researchers are therefore typically constrained to using controlled or model systems that are ethically, financially, and logistically feasible. For example, mice are often used to understand human disease, and quarantine greenhouses are used to study plant disease . These proxy systems are imperfect, as we intuitively know that there are differences between a mouse and a human, so there are differences between the potted plant on your windowsill and the towering redwoods of California. These differences are intuitive, but also biological as the effects of interactions between organisms and their complex environments are impossible to recreate, and these environmental and physiological differences directly affect the outcomes of research in plant pathology . The chances to experimentally infect plants in accurate conditions are few and far between, so rare opportunities provide critical research windows. One such system that suffers from a lack of realistic experimental control is Pierce’s disease of grapevines , a persistent burden to the vineyards of California since the introduction of the etiological agent into North America in the late 1800s . The pathogenic bacterial strains that colonized Vitis vinifera and cause PD have since spread to Taiwan, Europe, and the Eastern United States . Although the relationship between Xylella fastidiosa and PD has been understood since 1978 , this study is the first to document the progression of symptoms in mature commercial grapevines under field conditions over the course of several years with a known inoculation time and location on the plant. Despite the challenges of conducting this work in-situ, predominantly concern from the proprietors about infection spreading, we have made insights that counter much of the classic understanding of this disease system and deepen our understanding of symptom development and bacterial multiplication and movement in PD. Several studies have documented PD, which collectively create expectations for the progression of this disease. The pathogen is known to move through mature grapevines at least fast enough to enter the cordon from the shoot within the first year of infection . However, given that removal of the cordon does not reduce disease severity in subsequent years , the pathogen must be moving further into the plant, presumably into the trunk during the first year. It is also well established that there are differences in response to X. fastidiosa infection based on Vitis vinifera cultivar, with some cultivars more susceptible than others . Symptoms are presumed to first appear late in the season that infection occurred, except in especially susceptible cultivars, such as Chardonnay, in which they can arise sooner. Characteristic symptoms have been described as leaf scorch, uneven lignification of the shoots, shriveled berries, and “matchstick” petioles, a situation in which the leaf blade becomes detached from the petiole. In this system, disease symptoms and pathogen infection do not always persist through the winter, a phenomenon called overwinter curing. However, the mechanism for recovery has not yet been determined, and current explanations range from pathogen temperature susceptibility to plant defensive responses. Expectations for overwinter curing are both temperature and cultivardependent, and in Napa Valley, the expected recovery rate for early-season inoculations is around 30% . Progression of PD symptoms has been monitored under controlled conditions . However, such data may not translate to plants grown in the vineyard for various biotic and abiotic reasons. Greenhouse conditions do not account for the complex nutritional, pest, and other biotic  and abiotic pressures of the field , and greenhouse assays of grapevines are typically conducted with an excised shoot, rather than a mature plant. Observational studies are similarly limited because they often miss aspects such as recovery, asymptomatic infections, or unexpected symptoms by not relying on controls, but instead seeking out expected symptoms. These studies have given us a view on characteristics of PD and its progression that we were able to experimentally test.Three commercial vineyard blocks were selected as trial sites at two locations in Napa Valley. Plots were set up with permission from vineyard managers; however, the identity and specific locations of sites must be kept confidential. Each block was planted with one of three wine grape cultivars, Vitis vinifera cv. Cabernet Sauvignon, Chardonnay or Merlot.

While there are three distinctive subspecies of X. fastidiosa, and it would be desirable to be able to use those subspecies for management decisions, so far the subspecies have not been found to have sufficient resolution to define host range or to infer risk . Understanding the molecular basis of plant host specificity in X. fastidiosa is vital for predicting and acting upon host shifts, but these are processes yet to be described . Xylella fastidiosa is a member of the group Xanthomonadaceae, and phylogenetically clusters sister to Xanthomonas albilineans, technically within the paraphyletic genus Xanthomonas, although Xylella is considered a separate genus . Xylella spp., and Xanthomonas albilineans, are the only xylem limited Xanthomonadaceae and have convergently reduced genomes compared to the rest of the genus . Xylella also lacks a Type III Secretion system , a loss compared to its higher order taxonomic group. As the purpose of the T3SS in phytopathogens is to deliver effectors into living plant cells the loss has been hypothesized to be due to X. fastidiosa primarily interacting with non-living tissue; insect cuticle and mature xylem vessels . While the molecular basis of host range is not understood, cut flower bucket there are consistent patterns in the ability of particular X. fastidiosa isolates to infect specific plant hosts regardless of their environmental condition .

This implies that genetics, as opposed to only environmental conditions, underlie the relationship between isolates and plant hosts that allow for colonization. Recurring pathogen specificity to a particular host can be either explained through phylogenetic signal, where members of a clade have shared traits that allow for pathogenesis in that host, or by pathological convergence, where more distantly related strains have separately acquired mechanisms for virulence. Both processes have underlying genetics but each shows different phylogenetic patterns . Lastly, we have seen that deletion of rpfF, which controls cell-cell signaling via a diffusible signal factor , can expand the host range of X. fastidiosa . For example, removing the O-antigen from the exterior of X. fastidiosa cells allows the plant to quickly recognize X. fastidiosa and initiate immune responses, thus decreasing its likelihood of colonization of the plant . O-antigens are highly variable and evolve rapidly, and often are shown to have coevolutionary histories between symbiotic organisms as they are the first exposed part of any bacterium . In terms of phylogenetic methods, cophylogenies have shown no cospeciation between plant hosts and X. fastidiosa or any other congruence between the evolutionary histories of X. fastidiosa and its plant hosts . Based on the current data, it is not generally possible to tell if X. fastidiosa is undergoing host jumps or range expansions, however the data available so far suggests that both are occurring given that in certain situations we see strains able to infect multiple hosts while in other situations we see multiple strains co-existing in nature but no cross infections of hosts .

Using the influx of whole genome data generated in the past several years, we searched the genomes of X. fastidiosa for correlations with plant host species. The first method we pursued was conducting ancestral state reconstructions. Ancestral state reconstructions use genetic data , with a known phenotype for each taxon, to characterize the most likely state that each ancestral node of the tree would have possessed for the phenotype of interest. This tool has been used to understand host-pathogen interactions via ancestral state reconstructions in fungi and trematodes parasite systems . Ideally, we would be able to ask: what was the most likely ancestral host of the ancestor of all X. fastidiosa? If we can understand patterns in the past, it can help us better build models to predict future hosts based on the genomic changes associated with historical host shifts. Following the ancestral state reconstructions, we looked further into the pan-genome by calculating correlations between plant host types and the presence/absence of each gene. This study aimed to compare the commonly used genetic datasets available for phylogenetic analyses of X. fastidiosa both to compare phylogenetic topologies as well as ancestral host states from each dataset. We hypothesized that the pathogen phylogeny would be correlated with hosthistory and that we could observe this trend through ancestral state reconstruction. If there is no relationship between host and the phylogeny, there should not be conclusive ancestral state reconstruction results. We hypothesize that by using either the core genome of X. fastidiosa, pangenome phylogenetic tree, or both, it would be possible to estimate the likelihood of hypothetical plant hosts for ancestral nodes of inTherest .

This would show that the host is largely dependent and predictable based on the phylogeny of bacterial relationships and would lead to further pursuing allelic differences in core-genome and/or gene gain/loss in the pan-genome and estimate how either or both are correlated with plant host identity. While not biologically meaningful, since MLST data is still frequently used in X. fastidiosa management, we included that datatype in our analysis for comparison as well.Phylogenetic reconstruction of disparate regions and sizes are topologically similar The pan-genome of all sequences and the outgroup X. taiwanensis – Wufong1_PLS229 contained 17,024 genes . The alignment of MLST genes totaled 4,146 bp in length, while the core genome comprised 1,411 concatenated regions in a total of 354,816 bp. Non-recombinant regions identified with ClonalFrameML comprised only 32% of the core genome , leaving an alignment consisting of only 112,819 base pairs . The alignment contained 130 pairs of sequences that were completely identical to each other, highly reducing the amount of within subspecies differentiation that is possible with this dataset and creating large polytomies of indistinguishable sequences within subspecies fastidiosa as well as within subsp. pauca . Due to this lack of within-subspecies resolution, the phylogeny with recombinant regions removed is only suitable for between subspecies comparisons due to the extensive data loss in removing recombinant regions. The strains and locations in the alignment with recombination can be visualized in the supplemental materials . While between subspecies topologies are similar among the four trees generated, they are not identical. The core genome tree shows consensus of taxonomic division into three subspecies, however, subsp. sandyi and morus could be either part of subsp. fastidiosa or each their own small subspecies without affecting the monophyly of subsp. fastidiosa. . The non-recombinant tree is similar except that subsp. morus is clusThered within subsp. fastidiosa. The pan-genome splits the most basal of the three subspecies, subsp. pauca, into a paraphyletic cluster, however places multiplex, fastidiosa, morus, and sandyi similarly to the core phylogeny . The MLST tree shows subsp. morus as the outgroup to subsp. fastidiosa while subsp. sandyi falls within subsp. fastidiosa. The other difference among the four topologically similar trees is variation in branch length. Since the pangenome tree was built with gene presence/absence data, it was calculated in gene changes per site. Phylogeny and alignment information is summarized in Table 1. A 16S rDNA phylogeny was also built as a comparison , but the phylogeny provided very poor differentiation among strains .Within subspecies fastidiosa, the core genome rearticulates the three PD clades that were found in Castillo et al. . Within the clade defined as PD-III, the sequence similarity in the core has led to extensive polytomies, with many sequences indistinguishable in the core . The three PD clades are also articulated in the non-recombinant phylogeny, and the pan-genome phylogeny, however the MLST tree does not differentiate these clades from one another. Not poorly resolved, the MLST does have high bootstrap support for clades that conflict with trees constructed with core and pan genome trees, flower display buckets suggesting that using MLST genes has the potential to subvert the analysis of relationships between taxa, while showing strong bootstrap support. Within subspecies multiplex, there have typically been considered two groups, the non-recombining “non-IHR”, multiplex, as well as the recombining outgroup “IHR” multiplex .

The core genome tree as well as the MLST tree both articulate these two groups, the clade “non-IHR”, as well as the non-monophyletic recombining group, “IHR”. The non-recombinant tree and the pan-genome tree do not re-create these groupings . All phylogenies but the pan genome show a consistent split in subsp. pauca between the strains isolated from the Italian OQDS outbreak and the mixed host strains from Brazil. Within the OQDS strains, as well as several very closely related strains from Costa Rica, there is no clear resolution at this genomic scale. Within the Brazilian clade, strain Hib4 is the outgroup in all phylogenies except the MLST. subsp.Interrogating the results of the ancestral state reconstruction to the genus level of the core-genome phylogeny shows undetermined hosts at the deepest nodes . However, the ancestral node of the subspecies fastidiosa has a significant association with the plant genus Coffea, which persists throughout subspecies fastidiosa as the most likely ancestral host for all strains isolated from South and Central America. This changes for the Pierce’s disease of grapevines clade, where the ancestral host of all nodes except one is Vitis, the one exception being an ancestral Prunus node. Subspecies sandyi and morus are undetermined in ancestral hosts. Subspecies multiplex has a more dynamic history, with Vaccinium shown to be the most likely ancestral host for the subspecies, and then within the clade, a switch to a large group of nodes whose most likely host in Prunus, as well as two nodes depicting Platanus and Olea. Subspecies pauca does not have a determined ancestral host of the whole subspecies, and internal nodes switch several times between Citrus and Coffea, and once to Olea. In terms of the genera across the reconstructions, while the deep nodes , are often undetermined, there is more resolution within subspecies . The node that is consistent across the four reconstructions is that there is a high likelihood of the genus Coffea being the ancestral host of the node representing the introduction of subsp. fastidiosa from Central to North America. The genus Vaccinium was predicted as the most likely ancestral host of subsp. multiplex in the core genome phylogeny, whereas in the nonrecombinant phylogeny, the ancestor of all but one strain of subsp. multiplex is the genus Prunus. All four trees agree upon the ancestor of the internal “non-IHR” multiplex clade being Prunus. In terms of the transition models chosen for each reconstruction, most trees had lower AIC scores when using the equal rates model with fewer parameters than the symmetrical rates model, the exception being for the pan-genome super order reconstruction having a lower AIC score with the symmetric model than the equal rates .At the node representing the ancestor of the species X. fastidiosa, both the nonrecombinant core and pan-genome phylogenies predict that the clade Rosid is the most likely ancestral host . The core and MLST phylogenies predict Asterid to be the ancestral host, but at lower likelihoods of 87 and 78%, respectively, which are visualized, along with all likelihoods under 95%, as undetermined . There is enough discordance between reconstructions a consistent pattern at this host depth is unlikely.Bacteria isolated from the genera Coffea and Vitis as well as the super-orders Asterid and Rosid have X. fastidiosa genes with which they are significantly correlated; totaling 30 genes . Ten of these 30 genes are significantly correlated with both Asterids and Rosids, with paired, opposite relationships . Some correlations are of significance due to elevated presence of the gene among strains found in a particular host, while most are significant due to an absence of particular genes in the host of interest. Since lineage-specific interdependencies are accounted for with the phylogeny, the correlated genes are representative of convergent processes, either evolutionarily or via lateral gene transfer, not shared ancestry by descent. Genes that are significant mark repeated non-vertical descent changes in the pan-genome of strains in convergent patterns specific to the hosts of interest. While most identified genes are hypothetical proteins, genes shown to be correlated with host were fitB_1 system, involved in in-host migration, vbhT , socA , and a HTH-type transcriptional regulator .In this paper, we show that there is a genetic basis to the host range of X. fastidiosa. We demonstrate that both the phylogeny and gene gain and loss in the pan-genome are connected to plant host of the diverse species X. fastidiosa, and that an Asterid of undetermined genus was the most likely ancestral plant host of X. fastidiosa.

Alleviation of thermal limitations on Pc growth rates increased the predicted Pc risk in an 470 000 km2 area for the high emissions scenario with a projected mean increase of 0.08. For the low emissions scenario, Pc risks increase over a 390 000 km2 area for the low emissions scenario, with an average 0.05 increase in the Pc risk. Overall, the total area where Pc risks exceed 0.5 declines to approximately 164 000 km2 or 13% of the study region for the B1 scenario. The decline in spatial extent under a high emissions scenario is actually slightly less pronounced, with the predicted Pc range being 170 000 km2 or 13.7% of the region. This occurs even though the area where there is sufficient water for high spring Pc activity declines more for the high emissions scenario than for a low emissions scenario. The results reflect the importance of a trade off between increased winter survival range and decreased spring water availability. At the scale of the Bay Area, future climate projections contain considerable uncertainty associated with climatic down scaling. Broadly, however, the B1 scenario represents a case with warmer spring temperatures but without considerable reduction in rainfall,while the A2 scenario represents a situation with reduced rainfall and warmer temperatures. As shown in Fig. 3, cut flower transport bucket the effects are quite striking: an increase in temperature without a large decrease in rainfall results in an increase in the area that is vulnerable to Pc by approximately 20% .

Conversely, a decline in rainfall and increase in temperature result in a large decline in Pc risk, to approximately 3800 km2 , about half of its contemporary range.The modeling study indicated four different controls on modeled Pc range, which interacted to generate complex spatial patterns of Pc response to projected climatic changes. The first control was the proportion of soil organic carbon, which provided a static template of areas in which Pc would not establish. The second control was winter temperatures which impede pathogen over-winter survival. The range over which the pathogen survived winter expanded sequentially from contemporary conditions to the B1 and then A2 climate scenarios. The third and fourth controls on pathogen range lay in spring temperature and rainfall which together controlled spring Pc activity. The implications of drier climatic conditions in the study area varied depending on the absolute local Pc risk in a given area. Thus, little change in Pc risk was predicted in wet regions such as the northern coast, where drier spring conditions under B1 and A2 climate futures were still wet enough to support Pc activity. In the driest limits of the current Pc range, contemporary rainfall was too limited to support Pc activity, and further drying of the climate under B1 and A2 scenarios did not alter the pathogen risk. In mesic regions, progressively dry springs reduced overall Pc risk under the B1 scenario, relative to contemporary conditions.

Despite further reductions in rainfall under the A2 scenario, however, only minor further decreases in Pc spring risk were predicted compared to the B1 scenario. This we attribute to increasing spring temperatures enhancing the rate of Pc expansion during spring, which increased disease risk over part of the model range. The spatial locations where warmer spring temperatures under the B1 and A2 scenarios caused increases in Pc risk were spatially separated from the locations where decreases in rainfall reduced Pc risk, so the overall pattern of Pc risk during spring under contemporary, B1 and A2 scenarios differs: the B1 case primarily reflects reductions in the contemporary area of high risk, while the A2 scenario continues this reduction but also leads to expanded regions of moderate Pc risk in the north-eastern part of the study region. The interactions of these controls lead to a decrease in Pc range under both B1 and A2 scenarios compared to contemporary conditions, but, surprisingly, less of a reduction in total Pc range under the more extreme warming scenario. The modeling study ignored several factors that impact dynamic Pc risk – that is, the risk of Pc developing in uninfected areas. It has not considered pathways of introduction of Pc into natural communities. Proximity to roads, streams, soil disturbance, innoculum sources, and high vehicle or pedestrian traffic are highly likely to impact these risks. Similarly, the diversity of Californian vegetation and its variable sensitivity to Pc has not been considered: higher resolution studies considering differentials of vegetation susceptibility could provide a far more nuanced picture of the spatial extent of Pc, and could, for instance, eliminate areas of risk in primarily urbanized environments in the San Francisco Bay Area.

However, in the absence of improved data regarding host–pathogen interactions for dominant vegetation types in the area, there is limited basis on which to make such a high-resolution suite of assumptions about vegetation susceptibility. Despite this missing information, the damped sensitivity of Pc range relative to changes in assumed host resistance across the region in conjunction with field observations confirming that a wide range of regional species are susceptible to Pc, suggests that the overall conclusions of the modeling study regarding climate sensitivity should be robust. Finally, we have assumed that the local environmental conditions experienced by Pc are determined solely by climate, which is not true in the irrigated agricultural areas in California’s Central and Imperial Valleys. In these regions, irrigation regimes that are sufficiently frequent to sustain high water potentials in the root zone would alleviate the environmental water stress, as indicated by the observation of Pc occurrence within the Imperial Valley. If water stress is alleviated in these regions, temperatures are warm enough in both winter and spring to support high levels of Pc activity. Thus, irrigated agricultural land should be considered at risk of Pc infestation under both present and future scenarios. Overall, the modeling analysis suggests that Pc has a much larger potential range in the US southwest than could be inferred purely from current observations of the locations of infected sites. At present, despite the high awareness of Phytophthora ramorum and sudden oak death in the US southwest, there is limited public awareness, policy or scientific attention given to Pc . Sources of Pc are known to exist in agricultural and nursery settings in California , and appear to be the source of Pc infection in native ecosystems in several cases . However, no current phytosanitation certification programs, protocols for reducing soil movement from infected to clean sites, or recognized successful spot treatment approaches to minimizing Pc spread are in place in California . While the number of native species in California that are susceptible to Pc remains unclear, estimates in comparably biodiverse southwestern Australia suggest that over 3000 of the 5700 indigenous plant species are susceptible . Pc has already demonstrated its potential to reshape plant communities and entire ecologies . In the absence of detailed information about host susceptibility in California, the large ranges that appear to be supported by contemporary climates in conjunction with the severe effects of Pc in comparable ecosystems point to a critical need to improve risk assessment, procona flower transport containers phytosanitation and awareness of Pc disease. Despite its relatively simple ecology, Pc nonetheless displays non-monotonic responses to climatic warming at regional and local scales, with spatially distinct regions having opposite trends in Pc risk. Although the modeling framework presented accounts for many different aspects of climate, other changes including carbon fertilization due to enhanced atmospheric [CO2], changing land use and ecological thresholds have been neglected, and linking disease risk models to ecosystem outcomes remains challenging. For example, in Western Australia a 40 year drying trend has reduced Pc activity, but with the effect of replacing Pc induced mortality with drought stress stress . Considering that Pc mortality is linked to periods of drought stress that follow wet periods in which Pc causes root damage , limitation of Pc range due to drying, while beneficial for limiting expansion of the pathogen, may come at the expense of increased mortality for infected ecosystems.

The idea that one should be eating healthy to stay healthy is not a debate. Numerous studies show how particular foods individualistically effect human health, but none thus far, to our knowledge, have investigated about the combined impact of a specific diet on the human body as a whole. It is critical for us to understand which kinds of things we should eat and the ways in which their collective consumption will impact our bodies. According to Dr. Tomas J. Carlson, a distinguished pediatrician and ethnobotany researcher, choosing foods from every color in the rainbow is the key to good health. Each fruit and vegetable gets its natural color from the chemical composition of the exclusive phytonutrient in it. Interestingly, the presence of one molecule in one fruit/ vegetable does not necessarily reflect the same color in another type of fresh produce. For instance, although the rich red color in most red fruits and vegetables is naturally derived from the phytonutrient lycopene, most berries such as strawberries and raspberries do not contain lycopene. Instead, they contain brightly colored chemicals called anthocyanins, which are made in plants during ripening season through the joining of a molecule of a sugar with a molecule of their colorless “anthocyanidin” precursors. Anthocyanins are also found in raspberries, which are high in dietary fiber and vitamin C and have a low glycemic index because they contain 6% fiber and only 4% sugar per total weight. Higher quantities of fiber in the fruit, when consumed, helps lower the levels of low-density lipoprotein or the ‘unhealthy’ cholesterol to enhance the functionality of our heart and potentially induce weight loss. The exact pigment that anthocyanins reflect is partly dependent on the variance in acidity or alkalinity in different plants. Because of the relatively high pH of the tissues in blueberry plants, these chemicals turn blue in color during the ripening process of the fruit. Recent research in the Journal of Nutrition suggests that the abundant antioxidant properties in wild blueberries contributes to the reduction in the development of such disorders as Alzheimer’s Dementia and cognitive loss. A type of antioxidants selectively found in yellow and orange colored foods are called cryptoxanthins. In a study conducted by Bovier et al., it is shown that the combination of the beta form of these carotenoids with other sources of nutrients such as lutein and zeaxanthin in carrots, oranges, and corn leads to improved visual processing speed with regular consumption in young healthy subjects.10, 18While green produce mainly derives its pigmentation from chlorophyll, its white counterparts get their natural color from anthoxanthins, favonoid pigments that exhibit antioxidant properties. Among green fruits and vegetables, broccoli stands apart as the most nutritious because of the special combination in which its 3 glucosinolate phytonutrients are found. This “dynamic trio” makes what are called Isothiocyanates , the detox-regulating molecules in broccoli that enhance vitamin A in the form of beta-carotene. Many recent studies claim that the antioxidants in ITCs not only regulate metabolism and cholesterol levels when consumed but also act as cancer chemopreventive phytochemicals. Fruits that are on the same level as broccoli with regards to health in the white-produce family are bananas. Japanese Scientists reveal that the high amounts of vitamin B6, manganese, potassium and fiber in the ripened versions of these fruits can help prevent high blood pressure, protect against atherosclerosis, and improve immunity levels in regular eaters. Despite an enormous amount of scientific knowledge and evidence for the overall beneficial effect of a single fruit/vegetable and/or phytonutrient at a time on human health, no study so far, to our knowledge, has been able to conclusively link the validity of these claims to the whole human body. This offers the opportunity for one to test the combined impact of eating a colorful diet on humans through a systematic study. The purpose of our investigation is to apply a more holistic approach to the study of how the human body is effected as a result of a diet that is composed of all the colors of the rainbow. In other words, in addition to exploring the individual food stuf’s role in improving health, we want to analyze the outcome of the regular incorporation of a whole pack of colorful foods into one’s meals.

They contrast markedly with the mean estimate of / 0.02 with an upper bound of 0.06 obtained for the non-IHR group . Recombination in clade A. As noted earlier, Nunney et al. proposed that the two clades A and B in the X. fastidiosa subsp. multiplex tree of Parker et al. corresponded to isolates that had been subject to IHR versus those largely free of IHR . However, this hypothesis is undermined by the absence of evidence of intersubspecific introgression from X. fastidiosa subsp. fastidiosa to X. fastidiosa subsp. multiplex in the analysis of clade A by Parker et al. . However, in our reanalysis of these data using the targeted introgression test , we found that out of the 9 loci, 6 showed statistically significant evidence of IHR in alleles found in the clade A isolates . Furthermore, in this reanalysis, all of the 6 clade A genotypes carried recombinant alleles in at least 4 of these loci, while none of the 9 clade B genotypes showed any similar evidence of IHR. As noted previously, the 6 recombinant group isolates common to both studies were all classified into group A . They encompassed 4 of the genotypes identified by Parker et al. : Almond2 , the Blueberry1 group , the Blueberry2 group , 25 liter pot plastic and the Ragweed group . They showed significant evidence of IHR at 4 or 5 of the 9 loci , so that in total they showed significant evidence of IHR in 6 or 7 out of the 17 genes examined.

Plant hosts of the recombinant group. Given the identification of a discrete recombinant subgroup within X. fastidiosa subsp. multiplex, we investigated the possibility that it corresponded with a shift in the plant hosts that were infected. Among the 33 isolates defining 9 STs, there were 5 plant hosts represented more than once: ST27 on almond , ST27 and ST40 on purple leaf plum , ST22, ST28, ST42, and ST58 on giant ragweed , ST32 on blackberry , and ST42 and ST43 on blueberry , to which we can add American elm by including the 4 additional clade A isolates from Parker et al. . Closer examination of X. fastidiosa subsp. multiplex isolates from these host plants, using the 143 isolates subjected to MLST , showed that almond and purple leaf plum isolates were recombinant types only about one-quarter of the time . The ratios in American elm from Washington, DC , and western ragweed from Texas were substantially higher , but it was only in blackberry , and blueberry that 100% of the isolates were recombinant. While the sample size for blackberry was very small, the isolates were geographically separate and defined a single sequence type that was found on no other host. ST32 differed from all other STs in the recombinant group, except ST41, in carrying the nonrecombinant pilU3 allele . Blueberry isolates were better represented and again were isolated from two different states, Florida and Georgia. The 7 isolates that we typed defined two STs: ST43, which was found in both Florida and Georgia and was unique to these blueberry isolates, and ST42, which was isolated in Georgia on blueberry but also in Texas on western ragweed .

Analysis of the recombinant group ofX. fastidiosa subsp. multiplex showed three important results. First, intersubspecific recombination was shown to have occurred in 50% of 8 loci scatThered throughout the genome that were chosen independently of the data . Second, it was shown that the donor of the introgressed sequence was X. fastidiosa subsp. fastidiosa, a subspecies introduced from Central America into the United States as a single strain . However, the introgressed sequence at two of the loci did not come from any of the X. fastidiosa subsp. fastidiosa genotypes that have been found in the United States. This result suggests that another introduction of X. fastidiosa subsp. fastidiosa must have occurred, an introduction that resulted in successful IHR, after which the donor genotype seems to have disappeared. This involvement of an unexpected X. fastidiosa subsp. fastidiosa strain supports the hypothesis that the members of the recombinant group share a single ancestral IHR event. Third, the hypothesis that IHR has facilitated a shift to new hosts is strongly supported by the example of blueberry, where 10 isolates have been typed and potentially supportedby the example of blackberry . A link between the shift to a novel plant host and homologous recombination has not been previously identified. Of course, the direct causation of this link can never be proved without knowledge of the genetic changes driving this shift. It can always be argued that the link is fortuitous and that one or more point mutations in the nonrecombined X. fastidiosa subsp. multiplex genome are causal in the host shift. First, both blueberry and blackberry are native to the United States, so if only a simple genetic change was required to infect these species, why did the native nonrecombinant X. fastidiosa subsp. multiplex apparently never acquire these changes? Second, a similar but even more extensive mixing of the genomes of X. fastidiosa subspp. fastidiosa and multiplex is found in the only form of X. fastidiosa that infects another U.S. native plant, mulberry . Furthermore, in other bacterial species, it has been demonstrated that recombination can drive rapid evolution, both in the laboratory and, in the case of Helicobacter pylori, in mice .

Similarly, McCarthy et al. concluded that lineages of Campylobacter jejuni in chickens versus cattle and sheep were able to shift host type, because rapid adaptation was facilitated by recombination with the resident host population. In the study by Nunney et al. , it was shown that the recombinant genotypes formed a well-defined group , demonstrating that intersubspecific homologous recombination was not randomly distributed across the X. fastidiosa subsp. multiplex isolates. This work was based on a survey of 143X. fastidiosa subsp. multiplex isolates using just 8 loci. There were 33 isolates that showed some evidence of IHR in at least 1 locus: all but 2 showed statistically significant evidence in at least 2 loci, while the remaining 110 showed no such evidence . The generality of this discrete group of recombinant forms was supported by our analysis presented here of the sequence data from 9 more loci sequenced by Parker et al. . These loci divided isolates into 2 groups that appeared to correspond to the recombinant and non-IHR groups, respectively , even though Parker et al. found no evidence of IHR. Upon reanalysis, we found statistically significant IHR in 6 of the 9 loci in the clade A data but no evidence of IHR in the clade B data. Clade A included 6 isolates that we had typed in the present study, and each of these showed evidence of IHR in 4 or 5 of the additional 9 loci. Thus, 25 litre plant pot in two independent samplings that together examined 17 loci, there was clear evidence of substantial genomewide IHR in the recombinant group iso-lates, amounting to 50% of the genes showing IHR across the MLST locis plus the pilU locus . The average was higher when based on the loci sequenced by Parker et al. ; however, this was probably biased upwards by the manner in which the loci were chosen . None of the IHR events in 6 of the 9 loci identified using the targeted introgression test, or in the case of complete introgression, a chi-square test, were detected by Parker et al. using PHI and the 9 tests implemented in RDP . This failure of the standard tests of recombination to detect IHR was previously noted by Nunney et al. , motivating the development of their introgression test. We examined the hypothesis that the recombinant group STs were derived from a single IHR event involving a X. fastidiosa subsp. multiplex recipient and an X. fastidiosa subsp. fastidiosa donor. The distribution of allelic differences among the recombinant STs was consistent with them all being derived from a single initial event, but a small number of other intersubspecific and intrasubspecific recombination events would also be needed . More importantly, the genotypes seen in the recombinant group can be accounted for entirely, or very nearly so, based on a single X. fastidiosa subsp. fastidiosa donor genotype. For example, the substantial variation in cysG can all be accounted for by an ancestral introgression of X. fastidiosa subsp. fastidiosa allele 12 followed by subsequent intrasubspecific recombination of X. fastidiosa subsp. multiplex sequence to form the other two alleles . In contrast, variation at pilU could be accounted for by a second donor contributing the X. fastidiosa subsp. fastidiosa pilU9 allele, but it could also have arisen by a single mutation in pilU1 unique to the recombinant group.

A possible single X. fastidiosa subsp. multiplex recipient genotype was also identified . This genotype is consistent with a known ST: setting cysG to allele 3 makes the recipient identical to ST45, which was sampled from the states of California, Kentucky, and Texas . Elsewhere, we consider a slightly different hypothesis regarding the origin of the recombinant group in which the donor and recipient subspecies are reversed—i.e., that it was derived from a single IHR event, but involving an X. fastidiosa subsp. multiplex donor and an X. fastidiosa subsp. fastidiosa recipient; however, apart from the role reversal, the conclusions are unalThered . The ancestral reconstruction allows us to consider the second question posed earlier: is the donor consistent with the X. fastidiosa subsp. fastidiosa genotypes found in the United States? The answer is very clearly “no.” The original donor carried cysG12 and holC19 . These alleles are both found in isolates from Central America, but no X. fastidiosa subsp. fastidiosa isolate found in the United States comes close to matching this criterion: the most similar U.S. ST has a 12-bp mismatch. There has been extensive sampling of X. fastidiosa subsp. fastidiosa within the United States, based on 85 isolates sampled across the United States from 15 different host plants . There is very little variation within X. fastidiosa subsp. fastidiosa isolates from the United States, consistent with the hypothesis that all X. fastidiosa subsp. fastidiosa isolates currently found in the United States are derived from a single strain introduced from Central America . Based on these data, we conclude that the X. fastidiosa subsp. fastidiosa donor was introduced into the United States from Central America and recombined with a native X. fastidiosa subsp. multiplex genotype similar to ST45; however, this donor lineage of X. fastidiosa subsp. fastidiosa was ultimately unsuccessful and died out. We can never conclusively prove the absence of this genotype from North America. However, X. fastidiosa has been extensively sampled from many plant species throughout the United States, and no isolates of X. fastidiosa subsp. fastidiosa have been found that carry alleles similar to the inferred donor alleles cysG12 and holC19 ; indeed all X. fastidiosa subsp. fastidiosa isolates so far found in the United States are consistent with the introduction into the United States of just a single genotype . The transient presence of the donor genotype is consistent with a single large-scale introgression event founding the recombinant group. This raises the possibility that conjugation might have been involved; however, if this was the case, the genomic DNA was broken into pieces prior to homologous recombination, since the data show short regions of recombination. The data from the MLST loci plus pilU show 7 significant recombination events , and 3 of them included at least one recombination breakpoint. Since these loci range in length from roughly 400 to 700 bp, this result would be consistent with an average recombination length of no more than a few kb, similar to the 2.6-kb average length observed by Nunney et al. in a comparison of two X. fastidiosa subsp. fastidiosa genomes . Similarly, the regions identified from the data of Parker et al. showed the same pattern, with a high proportion of recombination breakpoints identified within the sequenced loci . In this context, it is important to note that Rogers and Stenger have found a conjugative plasmid in X. fastidiosa. Furthermore, a high rate of transformation has been demonstrated in the lab , and it has been shown that this process can result in efficient recombination with only a few hundred bases of homologous sequence .

We included all structurally intact moss and lichen tissue in the live biomass category. This was determined by tugging gently on the brown part of the moss or lichen ramet; the part that broke off was determined to be litter. Large samples were chopped into small pieces, mixed, and then subsampled for fresh and dry weights. Tissues were then dried at 60C for 48 h or more before weighing. Above ground vascular net primary production was estimated as the sum of the current year‘s apical growth, including leaves and stems. We did not measure secondary growth for understory plants and thus our ANPP values represent an underestimate for shrubs where secondary growth is likely important, mainly Salix spp. and trees less than 1.37 m tall. Apical growth was defined as that produced from apical or intercalary meristems during the current growing season; it was calculated by summing the masses of all current year‘s leaf, stem, grow raspberries in a pot and reproductive tissues in the quadrat harvested. Harden and others measured moss production in these sites by measuring the apical growth of individual species and then scaling growth to the plot level with digital mapping. At each site, an average of ten 60 · 60 cm2 moss plots were arranged along greater than 100 m transects with plots spaced every 20–40 m.

Percent cover values for up to five dominant moss species within each plot were recorded in fall 2001 via digital photos, extensive field notation, and digitization with Arcview 8.0 software . Apical growth for each species within each plot was based on growth between June and September of 2001. Within each plot, 10 cm2 dense, generally single-species patches of moss were dyed with a fluorescent brightener in early June. Sprayed moss samples were harvested in late September using a coring device of known area and refrigerated until measurement. Apical growth of each ramet was measured individually under a black light using calipers and new growth was harvested, dried, and weighed to estimate per ramet production. The density of stems per m2 was determined from the % cover plots described above. Moss NPP per species was then estimated on a per plot basis as the mass of apical growth per ramet times the ramet density per unit area times the areal coverage. To validate this method of estimating moss NPP, Harden and others compared estimates of H. splendens productivity to estimates based on a morphological growth marker . They found that the fluorescent dye method underestimated H. splendens production relative to the morphological method, possibly due to an offset in the timing of harvest of the two methods. We have chosen to report the fluorescent dye methods here because these estimates likely represent the most conservative estimate of moss NPP.

Compared with mature tundra, the disturbed ground associated with bear digs generally has lower vascular plant species richness. In contrast, Shannon-Weaver diversity levels are indistinguishable between mature tundra and digs, suggesting that the lower overall plant cover in bear digs allows a more even distribution of the abundances of plant species that are present. The results of this direct comparison are intuitive, but grizzlies’ cumulative effect on community richness and structure is more appropriately thought of as marginal addition to the mature tundra community than as a contrast to it. At any given time, recognizable bear digs make up only a small fraction of the ground cover in mature tundra areas. In the northern portion of our study area, Christian found between 7 and 23 bear digs within sight of each kilometer-long transect through ground squirrel habitat, and grizzlies digging for glacier lily bulbs excavate no more than 6% of suitable meadow habitat in a given year in Glacier National Park . The results of our pairwise comparison show that overall, mature tundra alone has fewer plant species than mature tundra with bear digs. The difference in our plots was small but statistically significant. So, although mature tundra harbors more plant species at the landscape level than bear digs, a potentially important effect emerges at the smaller scales relevant to most plants and many animals.

Square meter by square meter, repeated samples from a landscape of mature tundra dotted with bear digs will exhibit greater species richness than samples from a landscape of pure bear digs but also greater richness than samples from purely undisturbed tundra. Although our data cannot define the optimum level of disturbance, they suggest strongly that moderate levels of bear digging will modestly enhance local species richness. In alpine tundra, grizzly bear digs undergo transient increases in plant species richness and diversity that develop over time as recolonization of the dig occurs. This result is consistent with other studies of mammalian disturbance. In desert , coastal prairie , and serpentine annual grasslands , digging mammals have significant impacts on species richness and diversity, with species numbers usually rising after a disturbance event. In some cases, richness and diversity appear to decline as digs undergo succession back to mature tundra. This pattern fits that predicted by the intermediate disturbance hypothesis , with highest diversity or richness at an intermediate stage in succession following the bear digging. We note this result with caution, however, because our data indicate relative richness and diversity of bear digs and because our proxy for dig age is inherently imprecise. As recolonization occurs, in any case, we have shown that species composition of bear digs remains very similar to that of the immediately surrounding tundra, implying that throughout the successional recovery of bear digs most colonization comes from nearby plants.In addition to changing community diversity, bear digs have differential impacts on plants with different life histories. In particular, these disturbances appear to favor plants with certain clonal growth strategies. The biological significance of these effects differs between the two plant communities we studied. In snowbed areas, all clonal types except the most dominant one were favored by digging, suggesting that the statistically significant reduction in above ground guerrilla species may release species of other clonal types from competition, leading to the observed increases in species richness and diversity. In mesic meadows, by contrast, species of the dominant below ground phalanx clonal type were actually favored by digging, while two of the less common clonal types, above ground guerrilla and phalanx species, were suppressed. In this case, it is possible that bear digs are frequent enough to prevent the dominance of above ground species, even in what appears to be mature tundra. Our observations suggest that grizzly bear digs are very concentrated in some areas, but we cannot speculate on the recurrence interval for bear digs in any given location. Ultimately, several of the changes in group dominance by plant species of different clonal types were statistically insignificant. Given the broad range of plant morphologies and strategies grouped within each of our coarsely classified clonal types, we expected to find more significant differences at the scale of individual species. In a study of short-grass prairie communities, for example, Platt showed that a small suite of ‘‘fugitive’’ plant species depended entirely upon badger disturbances for recruitment and reproduction. Unlike Platt, but in common with other studies of digging mammals , we did not find individual species that appeared to absolutely require bear digs for establishment or reproduction. In retrospect, this result is not surprising, given the large number of abiotic disturbances that affect alpine communities; avalanches, frost heaving, solifluction, best grow pots and freeze-thaw ground tears are all significant forces in shaping tundra communities . Indeed, although the species most dependent on bear digs in this study, Epilobium anagallidifolium, also occurred in mature tundra plots, it did so almost exclusively in the cracks caused by frost heaving and ground tears.

The degree to which bear digs’ effects on biotic communities resemble the effects of these other, abiotic disturbances is not well understood, but it seems clear that the ground disturbance associated with bear foraging is not wholly unique in this environment. Still, we found several plant species that showed a statistically significant difference in abundance between bear digs and mature tundra. Furthermore, both species dominant on bear digs were of the phalanx clonal type, while five of the six more common in mature tundra were guerrilla species. These results raise questions about the role of clonal growth form in mediating competitive interactions. Theoretical work has emphasized that differences in clonal form can mediate competitive interactions , and Boeken and Shachak’s work in the Negev desert documented various plant traits favored by disturbances, including seed size and dispersal characteristics. Given the low seedling recruitment rates and long life spans thought to be typical of most tundra species, we suggest that grizzly bear digs may provide critical microsites for reproduction and survival of phalanx-type plants that are at a competitive disadvantage relative to aggressively spreading guerrilla species in densely vegetated mature tundra. Clearly, though, these interactions occur within a diverse plant community where species with widely varying clonal forms interact on small spatial scales. This diversity is probably maintained by the interaction between a diversity of clonal growth forms in the context of numerous microand mesoscale disturbance events. Our observation that the second most common species in a study plot frequently differs in clonal form from the dominant species supports the possibility that the varying clonal forms common in this system may help to allow coexistence of species.The growth of the “critical zone” paradigm has added impetus to closer investigation of soil-plant atmosphere interactions in ecohydrology . This follows from work emphasizing the importance of vegetation in regulating the global terrestrial hydrological cycle, with transpiration being the dominant “green water” flux to the atmosphere compared to evaporation from soils and canopy interception in most environments . More locally, the role vegetation plays in partitioning precipitation into such “green water” fluxes and alternative “blue water” fluxes to groundwater and streamflow has increased inTherest in the feedbacks between vegetation growth and soil development in different geographical environments . The emerging consequences of climatic warming to changes in vegetation characteristics and the implications of land use alterations add further momentum to the need to understand where plants get their water from, and how water is partitioned and recycled in soil-plant systems . Stable isotopes in soil water and plant stem water have been invaluable tools in elucidating ecohydrological interactions over the past decade . Earlier work by Ehleringer and Dawson explained the isotope content of xylem water in trees in terms of potential plant water sources. Building on that, Brooks et al. showed that the isotope characteristics of xylem water did not always correspond to bulk soil water sources as plant xylem water was fractionated and offset relative to the global meteoric water line compared to mobile soil water, groundwater and stream flow signatures. This led to the “Two Water Worlds” hypothesis which speculated that plant water was drawn from a “pool” of water that was “ecohydrologically separated” from the sources of groundwater recharge and stream flow . Research at some sites has found similar patterns of ecohydrologic separation and suggested it may be a ubiquitous characteristic of plant-water systems . Others have found that differences between plant water and mobile water may be limited only to drier periods , or may be less evident in some soil-vegetation systems . Direct hypothesis testing of potential processes that may explain the difference between the isotopic composition of xylem water and that of potential water sources has been advanced by detailed experiments in controlled environments, often involving the use of Bayesian mixing models which assume all potential plant water sources have been sampled . However, as field data become increasingly available from critical zone studies, more exploratory, inferential approaches can be insightful in terms of quantifying the degree to which xylem water isotopes can or cannot be attributed to measured soil water sources . As this research field has progressed, it has become apparent that extraction of soil and plant waters for isotope analysis is beset with a number of methodological issues . Soil waters held under different tensions may have different isotopic characteristics: for example, freely moving water sampled by suction lysimeters often shows a much less marked evaporative fractionation signal than bulk soil waters dominated by less mobile storage extracted by cryogenic or equilibration methods .

One metric especially stands out —positive predictive value for D. melanogaster was nearly 95% . In other words, given that an 11-day egg-laying segment of this species is classified as terminal, there is nearly a 95% chance that this classification is correct. The value of this performance metric in this fruit fly species contrasts to that for the values from data on the tephritid species where it ranged between around 66% to just under 75% depending on species and whether the database was complete or censored.We believe that there were a number of important results from this study. The first was the revelation that daily egg-laying patterns in fruit flies can be used to distinguish periods in a female fly’s life that end in death from those that do not . Most of the previous analyses and models concerned with individual-level egg laying in fruit flies at or near the end of life focused primarily on classification of end-of-life stage such as senescent, terminal or retired stage. Our approach differed in that we made specific predictions that could be quantified according to performance metrics. A second important result was that the overall performance metrics such as precision, container raspberries accuracy and false positives and negatives were heavily species-dependent. This was because of between-species differences in the consistency of egg laying patterns in both the midlife and terminal segments.

Among the three species the best performing metrics including accuracy and precision were those produced using data on D. melanogaster. A third important outcome of our study was that it shed important light on the challenges inherent in predicting the timing of death. As we noted earlier, one reason for some of the low-performing metrics was that the egg-laying patterns indicative of flies that are at the end of their lives are also present in midlife. Consequently the segments in which these occur are misclassified as terminal rather than midlife. Another problem of misclassification occurred when flies either died at the height of their egg laying or very old flies produced larger numbers of eggs close to their day of death. In other words, patterns of egg laying that were more common in midlife than at or near the end of life. These types of misclassifications were rare in D. melanogaster but occurred around one third of the time in the Mexfly as was reflected in the much higher rates of false negatives for this species relative to the other two. A fourth important and what we consider to be the overarching outcome of our study was “proof-of-principle”—patterns of reproduction at the individual level contain information that can be used to predict impending death in fruit flies. This general finding is important since it suggests that there are likely identifiable individual-level age-specific life history data ranging from information on respiratory, metabolic and locomotor activity to sleep, circadian, and mating patterns that can be used singly or in various combinations to predict the timing of death, not only in fruit flies, but in a range of other organisms including in humans.

Indeed, the egg laying patterns we used to identify impending death may be the rudimentary equivalents of clinical prediction rules in humans used to calibrate the timing of deaths across age groups and between sexes. A unique strength of our study was in its multi-level comparisons involving two different fruit fly families , two different genera within one of these families and three different species overall . Our results demonstrated that the same independent variables could be used across families, genera and species to classify egg-laying segments ; but that the performance metrics such as accuracy and precision were family-, genus- and species-dependent. A caveat that applies to all studies like ours involving contingency tables is that performance metrics including PPV and NPV depend upon the relative numbers of positive and negative cases. Because the proportion of negatives influences the number of both false positives and true negatives, as the proportion of negatives increases, PPV decreases and NPV increases. For example, in disease epidemiology the if the prevalence of infection were low, the PPV might be very small but will be much larger in a population with moderate or high prevalence. So the fraction of true positives relative to false positives will be lower in the former hypothetical scenario than in the latter but in both cases the TPR remains the same.

This sample size-dependent relationship for some performance metrics in the contingency table was one of the primary reasons we used equal numbers of mid life and terminal segments for each fruit fly species rather than the number of mid-life segments possible for each species. For example, after a 10-day maturation period, flies that live 25, 50 and 100 days each have only a single 11-day terminal segment but 4, 29 and 68 possible mid-life 11- day segments, respectively. Because both the cohort sizes and life expectancies differed among the three fly species, the relative numbers of possible mid-life versus terminal 11-day segments also differed. Thus using different relative numbers of mid-life versus terminal segments for each species rather than equal numbers would have confounded comparisons. This perspective underscores the caution required in the interpretation of contingency table performance metrics such as sensitivity and specificity where one is much higher than theother. Indeed, this is the case for the Medfly and Mexfly in which sensitivity ranges from 40 to 60% for both but specificity is substantially higher for both species—i.e., 90 and 70%, respectively. This we conclude that our findings are robust but that the specific values of the performance metrics are subject aspects of the sample including size and details such as segment length. An entirely different aspect of this large concept and one we don’t address in this study is that of distinguishing between mid-life and terminal segments when, for example, age-specific reproductive patterns are altered by changing food quality modifying availability . We believe that research on the classification and timing of death in non-human species has the potential to provide important insights into approaches on similar research in humans. This includes research in contexts ranging from estimations of time to death after withdrawal of life-sustaining treatment in patients including for organ transplants , in nursing homes where 50% die within 3 years and over 30% need palliative care within a year , in hospice end-of-life care where one of the key criteria for entry and Medicare eligibility is that the patient has 6 months or less to live if the disease takes its natural course , and for physicians and their patients who are considering euthanasia for what is referred to as “dying on time” for dementia patients who may not wish to continuing living in advanced stages of any number of dementia-related diseases, but the symptoms conflict with the due care criteria . Lowbush “wild” blueberries are considered a nutrient-rich healthy food, due in large part to their exceptional phenolic content and antioxidant activity. Lowbush blueberries are particularly rich in anthocyanins and the anthocyanin profile is complex compared with other fruits. They contain five of the six anthocyanidins commonly found in nature , which can have three different sugar moieties attached as well as acyl groups such as acetyl-, malonyl-, or coumaryl- also attached to the sugar moieties. Blueberries are also rich in proanthocyanidins, chlorogenic acid, and flavonols. Diets rich in blueberries or their polyphenolic-rich extracts have been associated with lower cardiovascular risk, weight gain and metabolic syndrome, draining pots and neurological diseases . In addition, studies involving blueberries have identified polyphenolic-derived phenolic acids that improve cell differentiation and proliferation of osteoblasts in vitro and promote bone growth and limit bone loss in rodents. These health-promoting effects are due to a myriad of mechanisms associated with blueberry polyphenolics, including prevention of oxidative stress and inflammation, and vaso- and lipid modulation.

Many human studies reporting positive health outcomes have used freeze-dried wild blueberry powder, which is a natural source of concentrated polyphenolics. However, the freeze-dried WBB powder may be tart or astringent and not always palatable to consume. This can be problematic in feeding trials in children and adults. In our previous work, we developed five food products prepared with freeze-dried WBB powder that were evaluated for children’s acceptability and desire to eat . These results are useful in designing food products as well as menu items that could be used in clinical trials of WBB-rich diets. In addition to evaluating sensory properties, it is important to validate the storage stability of polyphenolics in these products, before use in clinical trials, to ensure that a consistent dose of polyphenolics can be maintained. Blueberry polyphenolics, especially anthocyanins, are unstable in various processed forms such as juices, jams, purees, and canned berries when stored at ambient temperature. Additionally, anthocyanins in freeze-dried WBB powder are susceptible to degradation when stored at ambient temperature with a reported half-life of 139 days at 25 ◦C. The mechanism responsible for loss of anthocyanins during storage is unknown, but anthocyanin losses are commonly accompanied by increased polymeric color values, suggesting that anthocyanins form polymers with proanthocyanidins. In addition to polymerization, many other factors can affect the stability of anthocyanins including exposure to elevated temperatures, light, oxygen, metals, sugars, and ascorbic acid. At present, refrigeration of blueberry products such as jam and juices is the best approach to mitigate polyphenolic losses during storage. This study was undertaken to determine the stability of anthocyanins, flavonols, chlorogenic acid, and percent polymeric color in five blueberry products prepared with freeze-dried WBB powder. Gummy, oatmeal bar, graham cracker cookie, and juice were stored at 21 ◦C and 4.4 ◦C and evaluated for anthocyanin, flavonol, and chlorogenic acid content and percent polymeric color over eight weeks of storage. An ice pop product stored at −20 ◦C was evaluated for its anthocyanin and chlorogenic acid content over eight weeks of storage.Samples of juice, ice pop, gummy, oatmeal bar, and graham cracker cookie, each containing 15 g of WBB powder per serving, were prepared and packaged as previously described. One serving of oatmeal bar, ice pop, and graham cracker cookie was equivalent to one piece each , a juice serving was 135 g, and a gummy serving was 7 pieces, or 113 g. The amount of 15 g of WBB powder used in product formulations was calculated and converted from previous animal studies to humans. This involved only the use of brief microwave heating to solubilize the ingredients in order to avoid thermal loss of phenolic compounds, but still obtain a ready-to-consume non-baked product. The blueberry juice and ice pop were prepared with an anthocyanin concentrate, previously extracted from the WBB powder. This procedure was used to produce juice and ice pop products with no particulates. The formulation was adjusted with water so the anthocyanin content ofthe products was equivalent to that found in 15 g of WBB powder per serving. The preparation and processing of the samples for the storage study were performed in two separate experiments, using the same sample of wild freeze-dried blueberries obtained from FutureCeuticals Inc. . The WBB powder was stored at 15.5 ◦C for four months between the two experiments. The samples from Experiment 1 were stored at 21 ◦C and the samples from Experiment 2 were stored at 4.4 ◦C. The ice pop products prepared in Experiment 1 were stored at −20 ◦C. Three samples of each packaged product were evaluated at time 0 and after 2, 4, 6, and 8 weeks of storage.Polyphenolics were extracted by homogenizing 5 g of WBB-containing food product or 1 g of WBB powder in 25 mL of extraction solution containing methanol/water/formic acid , to the smallest particle size using a Euro Turrax T18 Tissuemizer for 1 min. Homogenates were centrifuged for 5 min at 10,864 × g. The pellet was re-extracted two additional times with 25 mL of extraction solution and centrifuged for 5 min at 10,864 × g. The filtrates were pooled and adjusted to 100 mL with extraction solvent in a volumetric flask. Prior to HPLC analysis, 5 mL of extract were dried in a Thermo Savant Speed Vac Plus SC210A and reconstituted in 1 mL 5% formic acid in water. All samples were passed through 0.45 µm nylon syringe filters into 1 mL HPLC vials prior to HPLC analysis.

The division of cultures into those with shorter incubation times and longer ones persists across the CLR transformation, though the magnitude of the differences decreases . Notably, the results here, which should mathematically represent trends that are valid in Euclidean space, also correspond to trends from the PCA analysis of the relative abundances still in simplex space. The agreement between the PCA results from the untransformed and those from the transformed datasets indicates that these results are very unlikely to be mathematical artifacts. In fact, they seem to carry enough biological significance to persist through extensive mathematical and statistical manipulation, including rarefaction, relative abundance calculations, and logratio transformation. Results from PCoA, dot plot of abundances, and PCA together serve as exceptional evidence of the compositional changes in the temporal cultures as well as the compositional “neutrality” of plaque. They also point to the lack of host driven differences, implicating the possibility of a baseline in vitro culture that can be developed by manipulating incubation times. A clear difference between the untransformed and CLR-transformed PCA that is worth some exposition is the number of factors shown graphically and accounted for mathematically.

In the untransformed PCA, blueberry pot size the first two components account for almost 90% of the variation in the samples while in the CLR-transformed PCA, the first two components only account for 45% of the total variation. Untransformed PCA on the temporal samples attributes variations to mostly the Streptococcus, Veillonella, and Pseudomonas OTUs while CLR-transformed PCA includes contributions from the Actinobacillus and Acinetobacter OTUs, along with a number of other genera. Even in the most prevalent and abundant genera, the PCA on the untransformed dataonly accounts for two of the three more prominent Streptococcus OTUs and one of the two Veillonella OTUs. On the other hand, results from the CLR-transformed PCA are much more consistent with the dot plot. Regardless of whether PCA on relative abundance data led to spurious trends in the temporal cultures, it is clear that performing PCA on simplex spaces can obscure the observation of important factors. For these reasons, we elected to keep performing PCA on CLR-transformed relative abundances in subsequent experiments, in order to gain a more complete understanding of what contributes to compositional differences.The results from the analyses of the temporal cultures confirm that our culturing procedures helped avoid contamination.

We discovered that controls in the temporal cultures were much more difficult to keep clean than those in the preliminary cultures. This may be due to the physical proximity of the control wells to the culture wells, which enables accidental splashing during feeding and moving of the plates, though we took great care to avoid doing so. As in the preliminary experiments, OTUs in the controls fell entirely within the taxa expected from either the E. coli spike-in or from the cultures, with no sign of external contamination. In terms of diversity, the temporal samples seemed slightly higher than the preliminary samples , though by what quantitative extent we do not know. We may estimate the differences between the preliminary cultures and the temporal cultures by observing the differences between the 24-hour cultures and the more aged cultures from the temporal samples. This particular comparison shows that the more aged cultures contained different distribution of OTUs and more OTUs in some cases. The increased diversity with increased incubation times was very much expected, and the procedures we used in to generate the temporal cultures were capable of capturing early colonizers as well as some of the middle colonizers from the oral bacterial community. In terms of procedural differences between the preliminary and temporal experiments, we used rarefaction to standardize and make samples comparable to one another, as rarefaction was once again shown to be a defensible approach in the temporal experiments.

We also opted not to separate the liquid from the sedimentary cells in the temporal experiment because of the lack of biological significance of the liquid samples. During the preliminary experiments, we confirmed that the liquid samples did not contain aberrant OTUs, i.e. OTUs belonging to neither the human oral bacterial community nor to the spike-in E. coli, and that the liquid samples contained low biomass. These observations led us to believe that the liquid was simply an extension of the sedimented cells, not a saliva-like substance that helps inoculate the cultures, especially given the lack of liquid movement in the cultures at all. As for the composition of the temporal cultures, we observed a shift from Streptococcus OTUs to Veillonella OTUs followed by the rise of Prevotella OTUs in all three hosts. This observation corroborates previous research on the succession of microbes in the colonization of the human dental surface, where Streptococcus spp. serve as the earliest colonizers of dental enamel, followed by microbes from several other genera that include Veillonella and Prevotella. This order of succession of species is in large part explained by the attachment capabilities of streptococcal bacteria. The two most abundant streptococcal OTUs in this study have been shown to produce surface proteins that bind to salivary amylase, human extra parotid glycoprotein, salivary surface lectins, human salivary antigen C fibrillar glycoprotein, and Veillonella fibrillar antigen B. Clearly, organisms in these streptococcal OTUs can act as intermediaries between the salivary pellicle deposited on dental enamel and other bacteria, and our methods provided appropriate conditions for this temporal succession observed in humans as well as in other in vivo and in vitro models of the human oral microbiome. We may be able to observe finer details of temporal development if we increase the sampling frequency, i.e. add more time points to the 5 existing ones during the 168-hour interval. However, it is not clear how the increased aerobic exposure would affect the purity and the development of the cultures. We will likely have to introduce additional measures to avoid contamination should we choose to sample the cultures more frequently. The E. coli spike-in in the temporal experiments served as a qualitative approach for estimating the biomass of the controls. For controls not contaminated by oral bacterial OTUs, the Escherichia-Shigella OTU dominated the relative abundance as in preliminary experiments . Even in controls that had been internally contaminated, the Escherichia-Shigella OTU still occupied 50% or more of the total abundance until the 168-hour incubation. At this time, we have confirmed that up to 168 hours of incubation time, the culturing procedure does not introduce microbes that interfere greatly with the in vitro compositions; both the OTUs and the estimated biomass in the controls, relative to the E. coli spike-in, imply that the OTUs and their distributions in the cultures were likely minimally affected by any potential cross-contamination between the wells. In future experiments, plant raspberry in container we may quantify the biomass directed by weighing the cells from the controls and cultures to gain an understanding of how much cross-contamination in a single plate affects the culture compositions in that plate, though we may not need to do so with the E. coli spike-in. Since the original purpose of the E. coli spike-in – developing a quantitative internal standard that would allow us to estimate the number of cells of the OTUs in the in vitro bacterial community – has been shown to be infeasible, we will eliminate this step to improve procedural efficiency as well as increase the sequencing depths of the cultures. Interestingly, a somewhat recent paper details a quantitative approach that links the number of cells per gram to the differences in gut microbiota.

The authors use flow cytometry cell-counting and 16S copy number corrections to bypass issues stemming from the fixed depth of Illumina MiSeq 16S rRNA sequencing process. The results from this quantitative profiling experiment showed that microbial load is significantly associated with changes in the microbiota, underscoring the importance of the ability to quantify sample biomass. It may be useful to incorporate this particular quantitation approach into future experiments, though the differences in microbial load are likely to be much higher in direct fecal samples across different hosts than the differences generated by cellular growth in in vitro cultures inoculated by similar amount of host plaque. Another potential improvement, should we reinstitute the spike-in procedure, may be using a species or type of bacteria other than E. coli. Ideally, the spike-in should be absent from the human oral microbiome while simultaneously demonstrating extraction and amplification efficiencies similar to members of the oral microbiome.

Because E. coli, a Gram-negative bacterium, has been shown to extract more efficiently than Grampositive bacteria such as S. aureus , E. coli K12 ER2738 may be a non-ideal choice for spike-ins. In this respect, we may be able to use the phi X 174 DNA sequences employed by the UC Davis Host-Microbe Systems Biology Core, as the sequences from that bacteriophage have been shown not to interfere much with the sequencing processes of microbes. Another key quantification step for reinstituting a spike-in would be to compare extraction efficiencies across different Gram-positive species in the oral bacterial community so that we understand potential extraction bias. With that information, we may be able to devise corrective measures to estimate the number of cells in each OTU more accurately. For the temporal cultures, we did not perform an in-depth analysis of host plaque because the goal for this phase of the project was not to generate in vitro cultures with composition that exactly matched that of host plaque. The purpose of this phase was to establish a baseline community that contained organisms common to oral and dental plaque microbiome, with some mimicry of the organismal succession observed in hosts. In this case, though the composition of plaque is useful for making comparisons, the absolute plaque composition from any single host and the precise replication of host specific oral communities were not of central importance. Furthermore, we have been prepared for the fact that unless able to perfectly reproduce the natural conditions, anyin vitro culturing process inevitably selects for a subset of organisms from the microbial community in nature. The short incubation time of one week and an environment absent of host nutrient sources and immune systems in this model make it extremely unlikely that we would be able to replicate the formation of a mature oral or dental plaque community. However, should we look to cultivate communities with higher diversity, we can now build upon this simple yet robust methodology by varying and/or adding nutrients, and regularly re-inoculating the cultures to increase the presence of middle and later colonizers. Here, re-inoculation may be the key to mimicking host variation in the “base” in vitro cultures. In any case, our methodology seemed capable of establishing communities with compositional developments that mimicked in vivo bacterial successions as well as capturing host differences in a qualitative, limited manner. As a last comment on the bio-informatics analyses: As we continued to augment the statistical methods from the preliminary experiments, we adopted the centered-logratio transformation to make PCA a mathematically sound process. This transformation helped confirm the biological significance of the patterns observed in the temporal cultures. The comparison between Figures 24 and 25 shows that CLR helps correct some potential artifacts created by working with relative abundance data, i.e. working in simplex space. For instance, the differences between clusters of samples in the PCA plot before transformation were much more pronounced than those after transformation . Given that CLR has been shown to reduce false discovery rates substantially, the decrease in the effect size in the temporal experiments as a result of the CLR transformation is simultaneously expected and highly desirable. Hence, we will use this method as a tool and a check on patterns observed in PCoA of relative abundance data going forward.In the temporal experiments, we expanded upon the previously established culturing procedures by extending the incubation time and implementing a feeding regimen. Analysis wise, we were able to characterize the community composition of the temporal cultures by using the established, validated bio-informatics pipeline. We also gained the capability to scrutinize and correct the results from Principal Component Analysis on relative abundance datasets, with the addition of the centered-log-ratio transformation.

Another gap in the oral microbiome model research is the effect of preservation on oral microbiomes and laboratory oral cultures. The lack of research in this area is perhaps unsurprising given the newness of this subfield, though the paucity by no means implies that preservation of microbiomes is by unimportant, especially when we consider the potential of preserving oral communities for inoculating in vitro cultures or the potential of preserving oral samples for therapeutic applications. While preservation experiments have been extensively performed on single bacteria species, preservation of microbiome samples has only been studied recently, with the primary emphasis on the gut microbiome. Research on the gut microbiome has provided evidence that storing samples at various temperatures with different preservatives does not significantly impact the structures of the bacterial communities, though sampling method matters; that cryopreservation, especially using glycerol and inulin, greatly helps maintain the composition and functionality of artificial gut microbiota; that cryopreservation is less detrimental to sensitive strains than lyophilization; that intra- and inter-subject variability outweighs the technical variability, blueberries in containers including variations between sequencing runs and as a consequence of long-term storage; and that after long term storage, freeze-dried fecal samples could still be used for transplantation treatment of C. difficile infections.

However, there has not been much research on the effects of preserving an oral bacterial community and even less on oral bacterial communities generated in vitro. The importance of studying the effects of preserving an in vitro oral bacterial community is two-fold: First, to examine the effects of preservation on an in vitro oral bacterial community; and second, to investigate the feasibility of preserving a universal inoculum for seeding in vitro oral microbiome cultures, wherein the inoculum is extracted from as simple of a model as possible. A universal oral microbiome inoculum would help bring a great degree of control into the generation of in vitro oral microbial models, potentially in oral disease research where a healthy baseline oral community needs to be established.As mentioned above, what would benefit the investigation of oral microbiome members and their roles in health and disease – specifically, the correlations and relationships of the organisms, their effects on one another during development in both healthy and disease, and the effects of various conditions and agents on the community – is the ability to reliably study reproducible versions of this community of organisms. This ability necessitates the development of a complex, stable, and generalizable in vitro model community.

While the construction of lab models of the bacterial oral microbiota has received a fair amount of attention, and some such models have produced communities that resemble the human oral microbiome at least in membership if not in relative abundances, most such models still require specialized incubators and fastidious surfaces and exhibit somewhat poor repeatability across both technical replicates and biological replicates. There is clear a need for a model of the human oral bacterial microbiome that is relatively simple to cultivate, cost-effective, easy to maintain for longer growth, and reproducible to the extent allowable by statistical margins of error from technical replicates. To this end, we intended for our work to contribute to the foundation for this model, by constructing and characterizing an in vitro culture that encompasses the initial stages of the oral community development. We began by focusing on the dental plaque community, as it requires the least fastidious surface and can help refine the methodological details needed for more complex in vitro oral models. The goals of this project were sequential in nature. First, we investigated the feasibility of generating an oral bacterial community in vitro with the resources available to our particular group. Then, we examined the temporal behavior of the community to see how closely it adhered to the time evolution of other models and to the development of the human oral community in vivo.

Thirdly, we determined whether the community would be fitting to serve as the basis for the preservation experiment. Lastly, we developed an approach that would help elucidate the apparent correlations and covariances discovered among the microorganisms during the preservation experiment. It was our hope that at its conclusion, the project would have provided us with not only data on how a minimally complex multi-organism in vitro oral microbiome performs under basic culturing conditions, but also data on how such a community responds to preservation and subsequent propagation. This data would help determine the degree of representation and generalizability of our model, and help create a “baseline” inoculum for in vitro oral bacterial microbiome models.To investigate the feasibility of the project and establish foundational procedures, we performed preliminary experiments to cultivate a complex in vitro supragingival dental microbiome. The community requires plaque from healthy hosts to serve as inoculum; the model is based on previously published research on in vitro dental communities, specifically, communities generated in 24-well culture plates without the use of additional substrata, such as removable hydroxyapatite disks that partially mimic dental enamel. We chose this basis for the model for its versatility in the initial stages of community formation, facile operation, and low cost. The goal for this phase of the project was to test whether the on-site facilities and conditions would allow us to generate a community that reasonably resembles the initial colonization stages of the host oral community. Should the community exhibit reasonable resemblance, it could then serve as the basis upon which to build subsequent parts of the model. In addition to investigating the feasibility and quality of the procedure to generate a simplified community, we also aimed to assemble and test reliable tools for this project, including DNA extraction for sequencing, potential quantitative determination of different types of cells by spike-ins, and bio-informatics analysis. While DNA extraction in this phase of the project was ultimately performed by the sequencing center at UC Davis , we decided to examine the extraction efficiency of two different commercial kits that were most commonly used in microbiome research at the time, with the goal of bringing this part of the process into the lab in later phases of the project. In terms of quantitative determination of the number of cells, we used E. coli as “doping” or “spiking-in” cells. We chose an E. coli strain that was easy to grow and manipulate, such that we could roughly specify the number of live cells – approximated by colony-forming units – for spiking the dental plaque cultures after incubation. It was our goal to use the known number of E. coli cells to estimate the number of sequencing reads, and then to use the number of E. coli reads, the number of reads from other OTUs, and 16S copy numbers for all OTUs from sequencing to estimate the number or concentration of cells of dental plaque bacteria in the model communities. To achieve these goals, we needed a solid bio-informatics foundation using either the QIIME or mothur pipeline, planting blueberries in pots so we also examined the results from both pipelines to determine which one better suited our purposes.A single healthy volunteer host was used for dental plaque sample collection during this phase of the project.

The host was a 29-year-old female in good health with no systemic diseases or family history of systemic diseases. The host routinely practiced good dental hygiene and had no periodontal diseases, but did have two fillings on two back molar teeth. Sample collection took place under Protocols 3-18-0189 and 3-19-0119, approved by the UCSB Human Subjects Committee. Sample collection took place as follows: Supragingival plaque of molar teeth, cheek side, was obtained using a sterilized Gracey curette; on average, plaque was collected from 5 molar teeth. Prior to plaque collection, hosts had abstained from food, non-water liquids, and dental hygiene for 12 hours. Collected plaque was immediately suspended in sterile SHI medium, gently mixed, and divided equally among wells in a sterile 24-well plate such that each well received 1.98mL of the mixture. Prior to receiving sterile or inoculated medium, wells were preconditioned with clarified human saliva , which was supplied as frozen pooled fractions from healthy human volunteers and clarified in the lab. Clarification consisted of centrifuging defrosted saliva at 6,000 x g for 3 minutes, mixing with 1X PBS in a 1:1 ratio, and passing the mixture through 0.2µm filters. Conditioning was performed by adding 150µL clarified saliva to each well, allowing saliva to air dry without the plate lid at 37°C for 60 minutes, and sterilizing with short-wave UV light for 60 minutes.After sterile and inoculated medium had been dispensed into the wells, each well received 20µL of 0.5% sucrose prior to being covered with the well-plate lid. Controls and cultures were incubated in triplicates or quadruplicates to test for repeatability. Lidded well plates were place in a closed chamber , and the chamber was flushed three times with an anaerobic gas mixture consisting of 85% nitrogen, 10% hydrogen, and 5% carbon dioxide, and then filled to positive pressure with the same gas mixture. During flushing and filling, gas from the cylinder was passed to the chamber through a 0.2µm filter to minimize potential microbial contamination. The chamber was then placed in an incubator set at 37°C and allowed to incubate for 20 to 24 hours. Some cultures were also incubated for four days to test the efficacy of commercial DNA extraction kits. The approximately one-day cultures were not fed again during incubation; each well of the four-day cultures was fed with 20µL of 0.5% sucrose every 24 hours before harvesting, with reestablishment of the anaerobic atmosphere after each feeding. After incubation, we observed for bacterial cell presence by looking for sedimented cells in the wells. From the wells with clear presence of cells , we harvested the wells by first aspirating 1.0mL of the liquid above the sedimented cultures into sterile microfuge tubes, and then mixing and pipetting the rest of the well into separate sterile microfuge tubes. For wells with no substantial cell presence , we gently and thoroughly mixed the wells and aspirated 1.0mL of the well contents into sterile microfuge tubes. Three samples from each type – controls, liquid above, and sedimented culture – were kept free of E. coli cells. We will henceforth refer to the E. coli cells as spike-ins and refer to the mixing with E. coli cells as “spiking”. Other samples received spike-ins at specified volumetric ratios, detailed in the E. coli Standard Curve and Spike-In subsection below. For controls and liquids with no spike-ins, the aforementioned 1.0mL aliquot in microfuge tubes was flash-frozen in liquid nitrogen and kept at -80°C until processing at the UC Davis Host Microbe Systems Biology Core . For sedimented cultures with no spike-ins, the 1.0mL of the culture was pelleted by centrifugation at 13,000 x g, the supernatant discarded, and the pellet flash-frozen and kept at -80°C until processing at the HMSB. Constitution and processing of spiked samples are described in the subsection titled E.coli Standard Curve and Spike-In.To observe the layers of sedimented cells, we reserved wells for microscopy without disturbing the layers, by carefully aspirating the liquid above and adding a commercial stain, which consisted of 200µL of a SYTO 9 and propidium iodide mixture from the FilmTracer LIVE/DEAD Biofilm Viability Kit . Dyes were allowed to penetrate the cells at room temperature for 30 minutes. To determine whether the surface of the bottoms of the wells affected the apparent viability of the cultured cells, we incubated cells in two different types of culture plates: standard tissue-culture-treated and surface-modified . Standard tissue-culture-treated plates have a net-negative surface charge while surface modified plates contain a mixture of positive and negative surface charges. Two plates of each type were used for this part of the preliminary experiments. Controls and cultures in both types of plates received identical treatment in terms of the volume of media, concentration of nutrients, and duration of incubation time. Results were visualized by staining and fluorescence microscopy procedures as described above.To determine the membership and abundances of the bacteria in these preliminary cultures using HTS, we attempted to develop an internal cell-counting standard using the E. coli K12 ER2738 strain . To do so, we first constructed a standard curve of colony-forming units vs. optical density at 600nm for ER2738 cells in the growth phase. For this curve, we grew overnight cultures of ER2738 and measured their optical density.

We maintain the assumption of two applications of grass herbicides, although we increase the treated acreage to 80 and 60 percent with one application composed of 40 percent molinate and 40 percent thiobencarb and the other composed of propanil on 60 percent of the acreage. Broadleaf control was adjusted to one application of triclopyr on 45 percent of the total rice acreage. Material and application costs of the herbicides are updated as well using information provided by UCCE . Finally, all cash operations are assumed to be financed at a nominal interest rate of 10.51 percent in accordance with the UCCE sample-costs document . As such, any change in the cost structure directly affects interest on operating capital, though the magnitude tends to be small. Overall, these updates result in a per-acre cost increase of $17.69 over the 2001 cost study. To more accurately represent the current world rice market , we assume the market price per cwt at harvest is the average price from 1986 through 2002 of $6.50 with average yields at 80 cwt per planted acre.

Government payments are divided into two components: direct payments and counter cyclical income-support payments as described by USDA, ERS . In accordance with the 2002 Farm Bill, square pots direct payments are calculated at 85 percent of average yields at $2.35 per cwt. Williams et al. estimate that growers of approximately 95 percent of planted acres have received this payment in the past, so the total direct payments are multipliedby 0.95. Counter cyclical income-support payments are calculated using the ERS formula, which we can summarize as 85 percent of average yields at $1.65 per cwt. Incorporation of these changes results in a $28.01 increase in gross revenue per acre over the 2001 UCCE sample-costs study. The original and adjusted costs and returns per acre are reported in Table 2. Given the public nature of experimental data on LibertyLink® rice grown in California and the full cooperation of Bayer CropScience through phone interviews and email correspondence, we use this transgenic variety as the basis for our analysis . We assume a price for Liberty® herbicide of $60 per gallon16 and an application rate of 0.446 pounds of AI per acre [500 grams AI per hectare ] in accordance with the company’s projected label recommendations .

To fully represent the fact that weed infestations will differ across plots, scenarios for transgenic cultivation are presented for both one and two applications of the herbicide on 100 percent of the acreage.The latter result is a direct consequence of the cost differential between ground and air applications of herbicides; ground applications of glufosinate cost approximately $12 per acre while air applications range from $6 to $7.25 per acre . The savings in chemical costs, however, drive the overall cost savings associated with transgenic rice and are explained using the information provided in Table 3. While the price of glufosinate per pound of AI is greater than all of the chemicals under consideration with the exception of triclopyr, the application rate per pound of AI is only 6 to 13 percent of the average herbicide control system. This decreases the cost of herbicide materials per acre by almost 62 percent as shown in the last column Table 2. When these results are combined, net returns over operating costs increase in the range of $45.89 to $74.90 per acre depending on the herbicide application rate, or $0.57 to $0.94 per cwt. Thus, this baseline scenario, which assumes perfect substitutability between medium-grain transgenic LibertyLink® rice and conventional varieties in terms of market acceptance and yields, predicts considerable economic incentives for rice growers to adopt transgenic varieties with similar characteristics due to their increased profitability. It is important to recognize, however, that these results are based on average costs over the entire Sacramento Valley ricegrowing region and utilize aggregate data to estimate the conventional herbicide weed-management regime.

Individual growers, of course, will most likely differ in regime from these averages depending on the characteristics of the specific operation. Those growers with “superior” land, as defined by lower aggregate weed-management costs, would benefit the least from adoption of transgenic rice while those with marginal land or serious weed-resistance problems tend to benefit more from the herbicide-management cost savings offered by the transgenic system and are hence most likely to adopt. To further investigate these issues, the assumption of perfect chemical substitutability, which essentially drives the assumption of identical yields, can be relaxed. A severely infested plot with a large, resistant seed bank of watergrass or some other weed would likely experience yield increases with adoption of a transgenic control system. Such yield gains have been observed in practice for HT soybeans and HT canola in the range of 0 to 20 percent . However, yields are not necessarily guaranteed to increase for all plots. Under generally ideal conditions, a yield drag of between 5 and 10 percent for medium-grain cultivars of LibertyLink® rice relative to conventional varieties has been observed in California rice field trials. This is consistent with similar field trials of HT soybeans. Such losses would decrease revenues and would thus reduce the increased profitability of adoption of this new technology. Yield drag should not be an issue with most growers given the advanced, widespread state of weed resistance to currently licensed chemicals for rice weed control in the Sacramento Valley. However, it is important to note that, in the short run, a few producers could actually experience a slight yield drag if the new technology was adopted; this is not expected to persist in the long run. A fall in demand for California rice due to consumer concerns, coupled with increased supply as a result of productivity gains, could cause rice prices to decline over time and decreasing net returns in the presence of yield changes. Similarly, a price premium for non-transgenic rice varieties could erode net-returns differences between traditional and HT cultivars but benefit conventional rice producers. Decreased yields or prices for transgenic rice, ceteris paribus, would reduce the gross rents from the technology. Furthermore, the seller of the transgenic seed is likely to charge a premium of up to 60 percent of total per-acre seed costs, depending on the pricing structure of the technology. Roundup Ready® and Bt seed for commercially produced transgenic crops has historically been priced from 30 to 60 percent higher than non-transgenic varieties, and price premia for LibertyLink® corn seed range from 0 to 30 percent, although average chemical costs per acre are typically greater . Furthermore, growers will likely pay at least part of the burden of the fees assessed by the CRCA. Assuming that these effects are constant per cwt of output, they can all be represented as a unit increase in costs in terms of net returns. Increased unit costs of this form, ceteris paribus, would alter the distribution of the rents between stakeholders but not dissipate gross rents. As points of reference, base assumptions on price and yields are $6.50 per cwt and 80 cwt per acre, large plastic pots so gross revenues from sales of rice output are assumed to be $520. A price premium of $0.25 per cwt for conventional rice as compared to transgenic rice with no associated change in yields would thus have the equivalent effect on net returns to the grower of a fee of about $20 per acre. Note that changing output prices does not affect the cost structure of the average farm operation and, thus, there is a direct, linear relationship between net returns and price. To calculate the impact of these effects, simple subtraction of the product of the price change and yield from the baseline scenario is appropriate. On the other hand, both a technology fee and the CRCA assessments directly enter the cost structure and, as such, affect interest costs as well. Tables 4 and 5 lay out these effects.

A 30 to 60 percent technology fee, assuming a seeding rate of 1.5 cwt per acre and price of conventional seed of $14 per cwt, is equivalent to $6.30 to $12.60 per acre. Total fees assessed as a result of the CRCA would currently be $8.50 per acre at identical seeding rates and yields of 80 cwt per acre, although it is unlikely that 100 percent of these assessments would be passed to the grower. Table 4 assumes no pass-through to growers of the legislated fees while Table 5 assumes the maximum pass-through, thus bounding the estimates. Both conservatively assume two applications of glufosinate per growing season. Without the CRCA legislation, adoption of LibertyLink® rice is profitable for a technology fee of $6.30 regardless of any realistic yield assumptions and profitable at a technology fee of $12.60 per acre so long as yield drag is no greater than 8.9 percent . With zero yield gains, net returns per acre in this range of seed price premium increase by between 21 and 25 percent over conventional rice returns with even greater benefits for those experiencing positive yield gains. If we assume a small price premium of, say, $0.25 per cwt, the technology is profitable for either yield losses of 7 percent with no technology fee or no yield change with an unrealistic $25.89 technology fee. This highlights the importance of yield and price assumptions on the calculation of net benefits. However, it is clear that, even with a small output price premium and a seed price premium at the upper end of the observed range, the most likely adopters will benefit from increased returns over costs. Allocation of maximum CRCA assessments to the grower slightly changes the per-acre benefits but does not affect the qualitative conclusions . Net returns over the baseline scenario with a $6.30 technology fee are no longer positive with an 8.6 percent yield drag nor for a $12.60 technology fee and a 6.7 percent yield drag. However, identical yields still result in net benefits of between $24.50 and $30.80 per acre, more than enough to cover a $0.25 price premium for conventional rice. To bound the per-acre benefits, we assume a lower bound of $0.25 per cwt price premium and an upper bound of no price premium with no CRCA pass-through. Under these assumptions, we conclude that the per-acre benefits of the transgenic HT technology are between –$7.22 and $58.10 for any given California rice grower with a midpoint estimate of $21.90. However, if we restrict attention to those producers most likely to adopt, as defined by at least zero difference in net returns, yield drag at the lower end of the range can be as high as 1.2 percent and they will still adopt.The preceding deterministic sensitivity analysis accounts for heterogeneity in land, weed infestation, and management ability as well as for the distribution of the rents generated by the technology. However, the magnitude of these rents is determined primarily through assumptions regarding yield and the price of rice as well as base assumptions on the price of alternative herbicide systems. While these point estimates are based on the best information available, another approach is to parameterize the distributions of those variables, which can be perceived as stochastic, and use Monte Carlo simulations to estimate the distribution of the surplus benefits of the transgenic rice technology. We take the specification in the equation and estimate distributions for a transgenic yield premium, the transgenic-rice price, and a conventional-rice price premium. Yields for the HT cultivar are assumed to vary according to a symmetric triangular distribution centered around 80 cwt per acre with a minimum value of 72 cwt and a maximum value of 88 cwt . This distribution allows for the possibility of yield gains and losses and, with symmetry, tends to be very conservative given the state of weed infestation and resistance across the state. Prices for California rice are essentially determined on the world market and thus are not influenced by the individual producer. Using historical data from USDA for 1986 through 2002, we assume a lognormal distribution for output price with a mean of $6.50 per cwt and a standard deviation of 1.67. Finally, the price premium for conventional rice is assumed to be distributed as a skewed triangular with a most-likely value of $0.25 , a minimum value of zero, and a maximum value of $0.52 or about 8 percent. These values are consistent with experience with corn, soybeans, and canola cited previously .

The system is divided into two parts. The first part is focused on storage and transmission of the EEG data for two different events: when the subject’s eyes are open and when eyes are closed. From researching brain waves, it was discovered that brain activity would be different between the two events. After careful study and testing, the conclusion was drawn that when the eyes are open, the Alpha and Beta wave features which are generated from the raw EEG signal are high. When the eyes are closed, Alpha and Beta features were low. However, these features were not enough to give reliable and appropriate results for the classification of these two events. Therefore, we tried to use different features, adding more features to the system to increase the reliability of the system. After working with one feature and two features, we extended the system to work with four different brain wavefeatures in the range of 0.5 to 30 with a sampling rate of 512. These features are Delta, Theta, Alpha and Beta.An Edge class is a powerful WuClass that has internal data storage and time series data processing capability. Like every WuClass, an edge class should be assigned a unique WuClass ID, which should be refered to in WuKong Standard Library.xml in master. In the EEGEdge Class, square pots the ID is explicitly defined by the WuClass annotation, so that the object instance of EEGEdgeClass can be discovered by the master.

According to the property declaration of WuClass in WuKong Standard Library.xml, a developer needs to define inputs and outputs for the edge class, and use the annotations for theedge framework to help developers to declare what kind of data structure they want to use to temporarily store upcoming time series data. The input property is declared as a buffer , whose data ring capacity is 2000 data points, index ring capacity is 30 units, and index is built every 1000 miliseconds. Therefore, the buffer will hold data in a time window of 30 seconds; it will at most keep 2000 data points. The buffer will store raw signal from EEG Device from which time series operators will fetch data and generate features. For every output property, an Edge class needs to define the corresponding set function. So, that a written WKPF property message will be generated and sent out when its value is updated. In the example, output is a property of output, whose property is 1. After defining properties as inputs and ouputs of the intelligent component, we implement two important methods of Edge Class. The register extension function will return a list of extensions, each of which is a handler of a specific stage in the data processing pipeline. In the EEG Edge Class, we registered EEG Feature Extractin Extension and EEG Execution Extension, both of which will be introduced in detail later. Beside providing data processing pipeline, the edge server also provides the communication pub/sub capability, so that an offline model can be initialized and updated through a pub/sub topic. model is loaded locally, so that there is no topic needed.Since models are offline trained in the EEG study, we ignore the online learning extension, and only focus on how to use models to do online classfication on the features extracted. Here, the EEG Execution Extension implements both Executable Number and Initiable.

Within the init function, we firstly load the model from local file system. The execute function accepts features in a list as the first parameters, and execution context as second parameter. Within the function, we use the model to classify whether we should label the features as eyes close or eyes open. Once we can know the probability of both eye close and eye open, we set 0.88 as the probability threshold to really trigger eyes close action by setting the output value. We tested the application on real physical devices.The importance of emotions lies in every-day-human-to-human communication and interaction. Understanding and recognizing human emotional status play an important role in human communication. Human-computer interface can play the same role of the human being to understand and recognize human emotions, and adjust its setting to fit with their human emotions. There are different approaches detecting and recognizing human emotions. First, facial expression is one of the earliest techniques used to detect human emotions and voice recognition based on the voice tone can detect emotions. However, these techniques are susceptible to deception, and vary from one situation to another. Second, Physiological signals are also used to detect emotion such as electrocardiogram and respiration. This approach provides more complex information than what is needed to detect emotion. Third, brain wave signal is used to detect emotion such as electroencephalogram , Electrocorticography and Functional magnetic resonance imaging . The advantages of using brain wave are its strong relevance for emotions and that it is not proneto deception.There is a spectrum of emotions which can differ from one person to another. There are different models presented by the research community to model human emotions. One model presents the basic emotions as happiness and sadness, another model presents the basic emotions as fear, anger, depression and satisfaction. Another model uses multiple dimensions or scales for emotion categorization. In this model emotion characterize by two main dimensions, valence and arousal.

Valence emotions range from positive to negative whereas arousal emotions range from low to high. For instance, fear can be defined as a negative valence and high arousal, whereas excitement can be defined as a positive valence, and high arousal. Figure shows the dimensional model of emotions.There are different type of stimuli to trigger emotion: self-eliciting, recalling, and using external stimulus. In order to stimulate the subject’s emotion in this thesis, I used excerpt videos from Liris-accede library. LIRIS-ACCEDE library has around 9800 good quality video excerpts used to induce different types of human emotions. This thesis used the library videos in the first trial and in the second trial. The third trial used a different resource to trigger emotions through videos that has marketing background. Only two emotions captured in the third trial which are High valence-High arousal, Low-valence and Low arousal. Figure shows some images from these different video excerpts.The aim of this project is to investigate the ability of a low-quality cheap commercial EEG Headset to classify different mental tasks. The EEG signal in this thesis was acquired using a single electrode placed on the forehead. The EEG signal was sent to the computer via Bluetooth. In order to integrate the EEG headset with the Wukong framework, large plastic pots the EEG Server Wuclass has been designed and built for the first two applications. Also, in order to build edge classification, the EEG edge WuClass has been designed and built for the third application.In the first two applications, this thesis intends to integrate the EEG headset with the Internet of Things framework ”WuKong” to allow the user to control external applications, such as turning lights on/off, playing music, etc. Two applications is built one application aimed to assist old people to do different tasks and the second application aimed to read a mind in an office. In the Old People Assistant Application, the system was built using two sensors. The first sensor is the camera, which captures eye blinking using the Haar algorithm from OpenCV library to classify and detect the eye blinking. The second sensor is the EEG headset which capture eSense signals, defined as attention and meditation signals and their values ranging from 0 to 100. Different patterns are built based on the attention or meditation signal values. In order to change from one state to another, or to keep on a certain state, the number of eye blinks has to be detected by the system. The result are shown in Chapter 3 for switching between 2 LEDs, but they can be extended to play different types of music and switch different LEDs, as shown in Figure 3.2. The second application is designed to show the mind state of a person inside in an office. In this application two WuClass have been built,the first is the EEG server WuClass which receives EEG signals, builds different patterns, and generates different index number for different pattern. The second Wulclass is the LED strips WuClass which receives the index number and plays the patterns that corresponds to the index number. In this application both eSense signals send to two EEG server in WuKong,then the WuKong framework has the ability to receive two different signals, attention and meditation.The third application is called classification eye states on the edge. In order to do this, the first step is to capture the raw EEG signal from EEG headset which done using a python code. The second step to build this application was the preprocessing steps that are applied to the collected EEG signals, including Fast Fourier Transform, removing EMG and EOG artifacts, and segment the EEG to one second segments.

The third step was to extract the EEG features, namely the power spectrum density features. In this application three different subjects volunteer to give their EEG signal, which acquired, preprocessed, and extracted the power spectrum density features. After collecting the EEG data we applied different machine learning algorithms in order to classify between two different eye states. We Used the accuracy of the classifier as our evaluation measurement. We reach 90% accuracy to classify between two different eye states using Support Vector Machine with Radial Basis Function kernel as our classifier. In order to use this classifier online in real time, we used the progression WuKong framework which supports intelligent edge. We built an EEG Edge class which has two components: first, feature extraction extension that uses relative intensive ration operator to calculate the intensive ration of EEG features; second, execution extension that loads the model which was generated offline, and based on the on the probability of the eye events, the threshold was chosen to be 0.88 to trigger an action when the eyes are closed.The emotions application presented in Chapter 4. The emotions application was implemented in order to classify between two different types of emotion. Using video clip from LIRIS-ACCEDE library which is designed to trigger these different emotion. Six subjects participated in this experiment which lasted for 60 seconds for high valence, high arousal and another 60 seconds for low valence, low arousal. Different types of features are extracted from the collected EEG signals, included time,frequency and nonlinear features, and 24 of these features are selected in order to use them with machine learning algorithms. Different machine algorithms are applied on the collected data, and the average accuracy using Linear discernment analysis was 79.31%. The model was evaluated using confusion matrix, which showed the diagonal is darker than the others, and those implies that the classifier is a good classifier . The second method used to evaluate the model was f1-score, which show an average of 83% for the subject number 4. The last method used to evaluate the model was Receiver Operating Characteristics which show the area under the curve with 91% for subject number 4. The last method used to evaluate the model was the learning curve in which the learning curve showed a misclassification range from 12% to 17% for the subject number 4.This thesis studies and discusses the ability of low-cost headset EEG to classify different eye events, and different mantel tasks such as attention, meditation and emotions. Table 5.1 compares different research literature that applied eyes events classification, using different EEG headset ranging from 4 – 14 electrodes with a cost of $799. Different features and classifiers are used to classify eye events. The accuracy ranges from 73% to 95%. The thesis presents a different approach to classify eyes events using low cost EEG headset that does not exceed $100. The power spectrum density features extracted , and Support Vector Machine with Radial basis function is used for classification. In this thesis, an accuracy of 90% is obtained to classify between eyes open and eyes close. With 90% accuracy of this work using only one electrode, and only cost of $99 is a better solution to detect the different eye states compared to other studies shown in Table 5.1.Table 5.2 shows data from research literature that discusses emotions classification using different EEG headsets.