A practical introduction to Random Forest for genetic association studies in ecology and evolution

Mitochondria, sex and variation in routine metabolic rate

Variation in the metabolic costs associated with organismal maintenance may play a key role in determining fitness, and thus these differences among individuals are likely to be subject to natural selection. Although the evolvability of maintenance metabolism depends on its underlying genetic architecture, relatively little is known about the nature of genetic variation that underlies this trait. To address this, we measured variation in routine metabolic rate (ṀO_{2 routine} ), an index of maintenance metabolism, within and among three populations of Atlantic killifish, Fundulus heteroclitus, including a population from a region of genetic admixture between two subspecies. Polygenic association tests among individuals from the admixed population identified 54 single nucleotide polymorphisms (SNPs) that were associated with ṀO_{2 routine} , and these SNPs accounted for 43% of interindividual variation in this trait. However, genetic associations with ṀO_{2 routine} involved different SNPs if females and males were analysed separately, and there was a sex-dependent effect of mitochondrial genotype on variation in routine metabolism. These results imply that there are sex-specific genetic mechanisms, and potential mitonuclear interactions, that underlie variation in ṀO_{2 routine} . Additionally, there was evidence for epistatic interactions between 17% of the possible pairs of trait-associated SNPs, suggesting that epistatic effects on ṀO_{2 routine} are common. These data demonstrate not only that phenotypic variation in this ecologically important trait has a polygenic basis with considerable epistasis among loci, but also that these underlying genetic mechanisms, and particularly the role of mitochondrial genotype, may be sex-specific.

Indications of strong adaptive population genetic structure in albacore tuna (Thunnus alalunga) in the southwest and central Pacific Ocean

Albacore tuna (Thunnus alalunga) has a distinctly complex life history in which juveniles and adults separate geographically but at times inhabit the same spaces sequentially. The species also migrates long distances and presumably experiences varied regimes of physical stress over a lifetime. There are, therefore, many opportunities for population structure to arise based on stochastic differences or environmental factors that promote local adaptation. However, with the extent of mobility consistently demonstrated by tagged individuals, there is also a strong argument for panmixia within an ocean basin. It is important to confirm such assumptions from a population genetics standpoint for this species in particular because albacore is one of the principal market tuna species that sustains massive global fisheries and yet is also a slow-growing temperate tuna. Consequently, we used 1,837 neutral SNP loci and 89 loci under potential selection to analyze population genetic structure among five sample groups collected from the western and central South Pacific. We found no evidence to challenge panmixia at neutral loci, but strong indications of structuring at adaptive loci. One population sample, from French Polynesia in 2004, was particularly differentiated. Unfortunately, the current study cannot infer whether the divergence is geographic or temporal, or possibly caused by sample distribution. We encourage future studies to include potentially adaptive loci and to continue fine-scale observations within an ocean basin, and not to assume genome-wide panmixia.

Comparing regression, naive Bayes, and random forest methods in the prediction of individual survival to second lactation in Holstein cattle

In this study, we compared multiple logistic regression, a linear method, to naive Bayes and random forest, 2 nonlinear machine-learning methods. We used all 3 methods to predict individual survival to second lactation in dairy heifers. The data set used for prediction contained 6,847 heifers born between January 2012 and June 2013, and had known survival outcomes. Each animal had 50 genomic estimated breeding values available at birth and up to 65 phenotypic variables that accumulated over time. Survival was predicted at 5 moments in life: at birth, at 18 mo, at first calving, at 6 wk after first calving, and at 200 d after first calving. The data sets were randomly split into 70% training and 30% testing sets to evaluate model performance for 20-fold validation. The methods were compared for accuracy, sensitivity, specificity, area under the curve (AUC) value, contrasts between groups for the prediction outcomes, and increase in surviving animals in a practical scenario. At birth and 18 mo, all methods had overlapping performance; no method significantly outperformed the other. At first calving, 6 wk after first calving, and 200 d after first calving, random forest and naive Bayes had overlapping performance, and both machine-learning methods outperformed multiple logistic regression. Overall, naive Bayes has the highest average AUC at all decision points up to 200 d after first calving. Random forest had the highest AUC at 200 d after first calving. All methods obtained similar increases in survival in the practical scenario. Despite this, the methods appeared to predict the survival of individual heifers differently. All methods improved over time, but the changes in mean model outcomes for surviving and non-surviving animals differed by method. Furthermore, the correlations of individual predictions between methods ranged from r = 0.417 to r = 0.700; the lowest correlations were at first calving for all methods. In short, all 3 methods were able to predict survival at a population level, because all methods improved survival in a practical scenario. However, depending on the method used, predictions for individual animals were quite different between methods.

Integrative Population and Physiological Genomics Reveals Mechanisms of Adaptation in Killifish

Adaptive divergence between marine and freshwater (FW) environments is important in generating phyletic diversity within fishes, but the genetic basis of this process remains poorly understood. Genome selection scans can identify adaptive loci, but incomplete knowledge of genotype-phenotype connections makes interpreting their significance difficult. In contrast, association mapping (genome-wide association mapping [GWAS], random forest [RF] analyses) links genotype to phenotype, but offer limited insight into the evolutionary forces shaping variation. Here, we combined GWAS, RF, and selection scans to identify loci important in adaptation to FW environments. We utilized FW-native and brackish water (BW)-native populations of Atlantic killifish (Fundulus heteroclitus) as well as a naturally admixed population between the two. We measured morphology and multiple physiological traits that differ between populations and may contribute to osmotic adaptation (salinity tolerance, hypoxia tolerance, metabolic rate, body shape) and used a reduced representation approach for genome-wide genotyping. Our results show patterns of population divergence in physiological capabilities that are consistent with local adaptation. Population genomic scans between BW-native and FW-native populations identified genomic regions evolving by natural selection, whereas association mapping revealed loci that contribute to variation for each trait. There was substantial overlap in the genomic regions putatively under selection and loci associated with phenotypic traits, particularly for salinity tolerance, suggesting that these regions and genes are important for adaptive divergence between BW and FW environments. Together, these data provide insight into the mechanisms that enable diversification of fishes across osmotic boundaries.

Genome-wide association study of an unusual dolphin mortality event reveals candidate genes for susceptibility and resistance to cetacean morbillivirus

Infectious diseases are significant demographic and evolutionary drivers of populations, but studies about the genetic basis of disease resistance and susceptibility are scarce in wildlife populations. Cetacean morbillivirus (CeMV) is a highly contagious disease that is increasing in both geographic distribution and incidence, causing unusual mortality events (UME) and killing tens of thousands of individuals across multiple cetacean species worldwide since the late 1980s. The largest CeMV outbreak in the Southern Hemisphere reported to date occurred in Australia in 2013, where it was a major factor in a UME, killing mainly young Indo-Pacific bottlenose dolphins (Tursiops aduncus). Using cases (nonsurvivors) and controls (putative survivors) from the most affected population, we carried out a genome-wide association study to identify candidate genes for resistance and susceptibility to CeMV. The genomic data set consisted of 278,147,988 sequence reads and 35,493 high-quality SNPs genotyped across 38 individuals. Association analyses found highly significant differences in allele and genotype frequencies among cases and controls at 65 SNPs, and Random Forests conservatively identified eight as candidates. Annotation of these SNPs identified five candidate genes (MAPK8, FBXW11, INADL, ANK3 and ACOX3) with functions associated with stress, pain and immune responses. Our findings provide the first insights into the genetic basis of host defence to this highly contagious disease, enabling the development of an applied evolutionary framework to monitor CeMV resistance across cetacean species. Biomarkers could now be established to assess potential risk factors associated with these genes in other CeMV-affected cetacean populations and species. These results could also possibly aid in the advancement of vaccines against morbilliviruses.

Tolerance traits related to climate change resilience are independent and polygenic

The resilience of organisms to climate change through adaptive evolution is dependent on the extent of genetically based variation in key phenotypic traits and the nature of genetic associations between them. For aquatic animals, upper thermal tolerance and hypoxia tolerance are likely to be a important determinants of sensitivity to climate change. To determine the genetic basis of these traits and to detect associations between them, we compared naturally occurring populations of two subspecies of Atlantic killifish, Fundulus heteroclitus, that differ in both thermal and hypoxia tolerance. Multilocus association mapping demonstrated that 47 and 35 single nucleotide polymorphisms (SNPs) explained 43.4% and 51.9% of variation in thermal and hypoxia tolerance, respectively, suggesting that genetic mechanisms underlie a substantial proportion of variation in each trait. However, no explanatory SNPs were shared between traits, and upper thermal tolerance varied approximately linearly with latitude, whereas hypoxia tolerance exhibited a steep phenotypic break across the contact zone between the subspecies. These results suggest that upper thermal tolerance and hypoxia tolerance are neither phenotypically correlated nor genetically associated, and thus that rates of adaptive change in these traits can be independently fine-tuned by natural selection. This modularity of important traits can underpin the evolvability of organisms to complex future environmental change.

Wild GWAS-association mapping in natural populations

The increasing affordability of sequencing and genotyping technologies has transformed the field of molecular ecology in recent decades. By correlating marker variants with trait variation using association analysis, large-scale genotyping and phenotyping of individuals from wild populations has enabled the identification of genomic regions that contribute to phenotypic differences among individuals. Such "gene mapping" studies are enabling us to better predict evolutionary potential and the ability of populations to adapt to challenges, such as changing environment. These studies are also allowing us to gain insight into the evolutionary processes maintaining variation in natural populations, to better understand genotype-by-environment and epistatic interactions and to track the dynamics of allele frequency change at loci contributing to traits under selection. Gene mapping in the wild using genomewide association scans (GWAS) do, however, come with a number of methodological challenges, not least the population structure in space and time inherent to natural populations. We here provide an overview of these challenges, summarize the exciting methodological advances and applications of association mapping in natural populations reported in this special issue and provide some guidelines for future "wild GWAS" research.