Climate-related naturally occurring epimutation and their roles in plant adaptation in A. thaliana

DNA methylation has been proposed to be an important mechanism that allows plants to respond to their environments sometimes entirely uncoupled from genetic variation. To understand the genetic basis, biological functions and climatic relationships of DNA methylation at a population scale in Arabidopsis thaliana, we performed a genome-wide association analysis with high-quality single nucleotide polymorphisms (SNPs), and found that ~56% on average, especially in the CHH sequence context (71%), of the differentially methylated regions (DMRs) are not tagged by SNPs. Among them, a total of 3235 DMRs are significantly associated with gene expressions and potentially heritable. 655 of the 3235 DMRs are associated with climatic variables, and we experimentally verified one of them, HEI10 (HUMAN ENHANCER OF CELL INVASION NO.10). Such epigenetic loci could be subjected to natural selection thereby affecting plant adaptation, and would be expected to be an indicator of accessions at risk. We therefore incorporated these climate-related DMRs into a gradient forest model, and found that the natural A. thaliana accessions in Southern Europe that may be most at risk under future climate change. Our findings highlight the importance of integrating DNA methylation that is independent of genetic variations, and climatic data to predict plants' vulnerability to future climate change.

Estimating resistance surfaces using gradient forest and allelic frequencies

Understanding landscape connectivity has become a global priority for mitigating the impact of landscape fragmentation on biodiversity. Connectivity methods that use link-based methods traditionally rely on relating pairwise genetic distance between individuals or demes to their landscape distance (e.g., geographic distance, cost distance). In this study, we present an alternative to conventional statistical approaches to refine cost surfaces by adapting the gradient forest approach to produce a resistance surface. Used in community ecology, gradient forest is an extension of random forest, and has been implemented in genomic studies to model species genetic offset under future climatic scenarios. By design, this adapted method, resGF, has the ability to handle multiple environmental predicators and is not subjected to traditional assumptions of linear models such as independence, normality and linearity. Using genetic simulations, resistance Gradient Forest (resGF) performance was compared to other published methods (maximum likelihood population effects model, random forest-based least-cost transect analysis and species distribution model). In univariate scenarios, resGF was able to distinguish the true surface contributing to genetic diversity among competing surfaces better than the compared methods. In multivariate scenarios, the gradient forest approach performed similarly to the other random forest-based approach using least-cost transect analysis but outperformed MLPE-based methods. Additionally, two worked examples are provided using two previously published data sets. This machine learning algorithm has the potential to improve our understanding of landscape connectivity and inform long-term biodiversity conservation strategies.

Using landscape genomics to delineate seed and breeding zones for lodgepole pine

Seed and breeding zones traditionally are delineated based on local adaptation of phenotypic traits associated with climate variables, an approach requiring long-term field experiments. In this study, we applied a landscape genomics approach to delineate seed and breeding zones for lodgepole pine. We used a gradient forest (GF) model to select environment-associated single nucleotide polymorphisms (SNPs) using three SNP datasets (full, neutral and candidate) and 20 climate variables for 1906 lodgepole pine (Pinus contorta) individuals in British Columbia and Alberta, Canada. The two GF models built with the full (28 954) and candidate (982) SNPs were compared. The GF models identified winter-related climate as major climatic factors driving genomic patterns of lodgepole pine's local adaptation. Based on the genomic gradients predicted by the full and candidate GF models, lodgepole pine distribution range in British Columbia and Alberta was delineated into six seed and breeding zones. Our approach is a novel and effective alternative to traditional common garden approaches for delineating seed and breeding zone, and could be applied to tree species lacking data from provenance trials or common garden experiments.

Developing Indicators of Nutrient Pollution in Streams Using 16S rRNA Gene Metabarcoding of Periphyton-Associated Bacteria

Indicators based on nutrient-biota relationships in streams can inform water quality restoration and protection programs. Bacterial assemblages could be particularly useful indicators of nutrient effects because they are species-rich, important contributors to ecosystem processes in streams, and responsive to rapidly changing conditions. Here, we sampled 25 streams weekly (12-14 times each) and used 16S rRNA gene metabarcoding of periphyton-associated bacteria to quantify the effects of total phosphorus (TP) and total nitrogen (TN). Threshold indicator taxa analysis identified assemblage-level changes and amplicon sequence variants (ASVs) that increased or decreased with increasing TP and TN concentrations (i.e., low P, high P, low N, and high N ASVs). Boosted regression trees confirmed that relative abundances of gene sequence reads for these four indicator groups were associated with nutrient concentrations. Gradient forest analysis complemented these results by using multiple predictors and random forest models for each ASV to identify portions of TP and TN gradients at which the greatest changes in assemblage structure occurred. Synthesized statistical results showed bacterial assemblage structure began changing at 24 μg TP/L with the greatest changes occurring from 110 to 195 μg/L. Changes in the bacterial assemblages associated with TN gradually occurred from 275 to 855 μg/L. Taxonomic and phylogenetic analyses showed that low nutrient ASVs were commonly Firmicutes, Verrucomicrobiota, Flavobacteriales, and Caulobacterales, Pseudomonadales, and Rhodobacterales of Proteobacteria, whereas other groups, such as Chitinophagales of Bacteroidota, and Burkholderiales, Rhizobiales, Sphingomonadales, and Steroidobacterales of Proteobacteria comprised the high nutrient ASVs. Overall, the responses of bacterial ASV indicators in this study highlight the utility of metabarcoding periphyton-associated bacteria for quantifying biotic responses to nutrient inputs in streams.

Adaptive significance of flowering time variation across natural seasonal environments in Arabidopsis thaliana

The relevance of flowering time variation and plasticity to climate adaptation requires a comprehensive empirical assessment. We investigated natural selection and the genetic architecture of flowering time in Arabidopsis through field experiments in Europe across multiple sites and seasons. We estimated selection for flowering time, plasticity and canalization. Loci associated with flowering time, plasticity and canalization by genome-wide association studies were tested for a geographic signature of climate adaptation. Selection favored early flowering and increased canalization, except at the northernmost site, but was rarely detected for plasticity. Genome-wide association studies revealed significant associations with flowering traits and supported a substantial polygenic inheritance. Alleles associated with late flowering, including functional FRIGIDA variants, were more common in regions experiencing high annual temperature variation. Flowering time plasticity to fall vs spring and summer environments was associated with GIGANTEA SUPPRESSOR 5, which promotes early flowering under decreasing day length and temperature. The finding that late flowering genotypes and alleles are associated with climate is evidence for past adaptation. Real-time phenotypic selection analysis, however, reveals pervasive contemporary selection for rapid flowering in agricultural settings across most of the species range. The response to this selection may involve genetic shifts in environmental cuing compared to the ancestral state.

Landscape analyses using eDNA metabarcoding and Earth observation predict community biodiversity in California

Ecosystems globally are under threat from ongoing anthropogenic environmental change. Effective conservation management requires more thorough biodiversity surveys that can reveal system-level patterns and that can be applied rapidly across space and time. Using modern ecological models and community science, we integrate environmental DNA and Earth observations to produce a time snapshot of regional biodiversity patterns and provide multi-scalar community-level characterization. We collected 278 samples in spring 2017 from coastal, shrub, and lowland forest sites in California, a complex ecosystem and biodiversity hotspot. We recovered 16,118 taxonomic entries from eDNA analyses and compiled associated traditional observations and environmental data to assess how well they predicted alpha, beta, and zeta diversity. We found that local habitat classification was diagnostic of community composition and distinct communities and organisms in different kingdoms are predicted by different environmental variables. Nonetheless, gradient forest models of 915 families recovered by eDNA analysis and using BIOCLIM variables, Sentinel-2 satellite data, human impact, and topographical features as predictors, explained 35% of the variance in community turnover. Elevation, sand percentage, and photosynthetic activities (NDVI32) were the top predictors. In addition to this signal of environmental filtering, we found a positive relationship between environmentally predicted families and their numbers of biotic interactions, suggesting environmental change could have a disproportionate effect on community networks. Together, these analyses show that coupling eDNA with environmental predictors including remote sensing data has capacity to test proposed Essential Biodiversity Variables and create new landscape biodiversity baselines that span the tree of life.

Landscape analyses using eDNA metabarcoding and Earth observation predict community biodiversity in California

Ecosystems globally are under threat from ongoing anthropogenic environmental change. Effective conservation management requires more thorough biodiversity surveys that can reveal system-level patterns and that can be applied rapidly across space and time. Using modern ecological models and community science, we integrate environmental DNA and Earth observations to produce a time snapshot of regional biodiversity patterns and provide multi-scalar community-level characterization. We collected 278 samples in spring 2017 from coastal, shrub, and lowland forest sites in California, a complex ecosystem and biodiversity hotspot. We recovered 16,118 taxonomic entries from eDNA analyses and compiled associated traditional observations and environmental data to assess how well they predicted alpha, beta, and zeta diversity. We found that local habitat classification was diagnostic of community composition and distinct communities and organisms in different kingdoms are predicted by different environmental variables. Nonetheless, gradient forest models of 915 families recovered by eDNA analysis and using BIOCLIM variables, Sentinel-2 satellite data, human impact, and topographical features as predictors, explained 35% of the variance in community turnover. Elevation, sand percentage, and photosynthetic activities (NDVI32) were the top predictors. In addition to this signal of environmental filtering, we found a positive relationship between environmentally predicted families and their numbers of biotic interactions, suggesting environmental change could have a disproportionate effect on community networks. Together, these analyses show that coupling eDNA with environmental predictors including remote sensing data has capacity to test proposed Essential Biodiversity Variables and create new landscape biodiversity baselines that span the tree of life.

Distinct interspecific and intraspecific vulnerability of coastal species to global change

Characterising and predicting species responses to anthropogenic global change is one of the key challenges in contemporary ecology and conservation. The sensitivity of marine species to climate change is increasingly being described with forecasted species distributions, yet these rarely account for population level processes such as genomic variation and local adaptation. This study compares inter- and intraspecific patterns of biological composition to determine how vulnerability to climate change, and its environmental drivers, vary across species and populations. We compare species trajectories for three ecologically important southern African marine invertebrates at two time points in the future, both at the species level, with correlative species distribution models, and at the population level, with gradient forest models. Reported range shifts are species-specific and include both predicted range gains and losses. Forecasted species responses to climate change are strongly influenced by changes in a suite of environmental variables, from sea surface salinity and sea surface temperature, to minimum air temperature. Our results further suggest a mismatch between future habitat suitability (where species can remain in their ecological niche) and genomic vulnerability (where populations retain their genomic composition), highlighting the inter- and intraspecific variability in species' sensitivity to global change. Overall, this study demonstrates the importance of considering species and population level climatic vulnerability when proactively managing coastal marine ecosystems in the Anthropocene.

DNA metabarcoding effectively quantifies diatom responses to nutrients in streams

Nutrient pollution from human activities remains a common problem facing stream ecosystems. Identifying ecological responses to phosphorus and nitrogen can inform decisions affecting the protection and management of streams and their watersheds. Diatoms are particularly useful because they are a highly diverse group of unicellular algae found in nearly all aquatic environments and are sensitive responders to increased nutrient concentrations. Here, we used DNA metabarcoding of stream diatoms as an approach to quantifying effects of total phosphorus (TP) and total nitrogen (TN). Threshold indicator taxa analysis (TITAN) identified operational taxonomic units (OTUs) that increased or decreased along TP and TN gradients along with nutrient concentrations at which assemblages had substantial changes in the occurrences and relative abundances of OTUs. Boosted regression trees showed that relative abundances of gene sequence reads for OTUs identified by TITAN as low P, high P, low N, or high N diatoms had strong relationships with nutrient concentrations, which provided support for potentially using these groups of diatoms as metrics in monitoring programs. Gradient forest analysis provided complementary information by characterizing multi-taxa assemblage change using multiple predictors and results from random forest models for each OTU. Collectively, these analyses showed that notable changes in diatom assemblage structure and OTUs began around 20 µg TP/L, low P diatoms decreased substantially and community change points occurred from 75 to 150 µg/L, and high P diatoms became increasingly dominant from 150 to 300 µg/L. Diatoms also responded to TN with large decreases in low N diatoms occurring from 280 to 525 µg TN/L and a transition to dominance by high N diatoms from 525-850 µg/L. These diatom responses to TP and TN could be used to inform protection efforts (i.e., anti-degradation) and management goals (i.e., nutrient reduction) in streams and watersheds. Our results add to the growing support for using diatom metabarcoding in monitoring programs.

Precipitation and vegetation shape patterns of genomic and craniometric variation in the central African rodent Praomys misonnei

Predicting species' capacity to respond to climate change is an essential first step in developing effective conservation strategies. However, conservation prioritization schemes rarely take evolutionary potential into account. Ecotones provide important opportunities for diversifying selection and may thus constitute reservoirs of standing variation, increasing the capacity for future adaptation. Here, we map patterns of environmentally associated genomic and craniometric variation in the central African rodent Praomys misonnei to identify areas with the greatest turnover in genomic composition. We also project patterns of environmentally associated genomic variation under future climate change scenarios to determine where populations may be under the greatest pressure to adapt. While precipitation gradients influence both genomic and craniometric variation, vegetation structure is also an important determinant of craniometric variation. Areas of elevated environmentally associated genomic and craniometric variation overlap with zones of rapid ecological transition underlining their importance as reservoirs of evolutionary potential. We also find that populations in the Sanaga river basin, central Cameroon and coastal Gabon are likely to be under the greatest pressure from climate change. Lastly, we make specific conservation recommendations on how to protect zones of high evolutionary potential and identify areas where populations may be the most susceptible to climate change.

Machine Learning for Health Services Researchers

BACKGROUND

Machine learning is increasingly used to predict healthcare outcomes, including cost, utilization, and quality.

OBJECTIVE

We provide a high-level overview of machine learning for healthcare outcomes researchers and decision makers.

METHODS

We introduce key concepts for understanding the application of machine learning methods to healthcare outcomes research. We first describe current standards to rigorously learn an estimator, which is an algorithm developed through machine learning to predict a particular outcome. We include steps for data preparation, estimator family selection, parameter learning, regularization, and evaluation. We then compare 3 of the most common machine learning methods: (1) decision tree methods that can be useful for identifying how different subpopulations experience different risks for an outcome; (2) deep learning methods that can identify complex nonlinear patterns or interactions between variables predictive of an outcome; and (3) ensemble methods that can improve predictive performance by combining multiple machine learning methods.

RESULTS

We demonstrate the application of common machine methods to a simulated insurance claims dataset. We specifically include statistical code in R and Python for the development and evaluation of estimators for predicting which patients are at heightened risk for hospitalization from ambulatory care-sensitive conditions.

CONCLUSIONS

Outcomes researchers should be aware of key standards for rigorously evaluating an estimator developed through machine learning approaches. Although multiple methods use machine learning concepts, different approaches are best suited for different research problems.

Applying landscape genomic tools to forest management and restoration of Hawaiian koa (Acacia koa) in a changing environment

Identifying and quantifying the importance of environmental variables in structuring population genetic variation can help inform management decisions for conservation, restoration, or reforestation purposes, in both current and future environmental conditions. Landscape genomics offers a powerful approach for understanding the environmental factors that currently associate with genetic variation, and given those associations, where populations may be most vulnerable under future environmental change. Here, we applied genotyping by sequencing to generate over 11,000 single nucleotide polymorphisms from 311 trees and then used nonlinear, multivariate environmental association methods to examine spatial genetic structure and its association with environmental variation in an ecologically and economically important tree species endemic to Hawaii, Acacia koa. Admixture and principal components analyses showed that trees from different islands are genetically distinct in general, with the exception of some genotypes that match other islands, likely as the result of recent translocations. Gradient forest and generalized dissimilarity models both revealed a strong association between genetic structure and mean annual rainfall. Utilizing a model for projected future climate on the island of Hawaii, we show that predicted changes in rainfall patterns may result in genetic offset, such that trees no longer may be genetically matched to their environment. These findings indicate that knowledge of current and future rainfall gradients can provide valuable information for the conservation of existing populations and also help refine seed transfer guidelines for reforestation or replanting of koa throughout the state.