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

Abstract

Large genomic studies are becoming increasingly common with advances in sequencing technology, and our ability to understand how genomic variation influences phenotypic variation between individuals has never been greater. The exploration of such relationships first requires the identification of associations between molecular markers and phenotypes. Here, we explore the use of Random Forest (RF), a powerful machine-learning algorithm, in genomic studies to discern loci underlying both discrete and quantitative traits, particularly when studying wild or nonmodel organisms. RF is becoming increasingly used in ecological and population genetics because, unlike traditional methods, it can efficiently analyse thousands of loci simultaneously and account for nonadditive interactions. However, understanding both the power and limitations of Random Forest is important for its proper implementation and the interpretation of results. We therefore provide a practical introduction to the algorithm and its use for identifying associations between molecular markers and phenotypes, discussing such topics as data limitations, algorithm initiation and optimization, as well as interpretation. We also provide short R tutorials as examples, with the aim of providing a guide to the implementation of the algorithm. Topics discussed here are intended to serve as an entry point for molecular ecologists interested in employing Random Forest to identify trait associations in genomic data sets.

The worldwide impact of urbanisation on avian functional diversity

“…We assessed the extent to which functional traits predict the tolerance of species to urbanisation with a Random Forests (RF) approach using the package randomForest (Liaw & Wiener 2002). RF is a machine‐learning algorithm that can efficiently analyse many predictors simultaneously and account for interactions (Brieuc et al 2018). In addition, we also modelled presence/absence of species in the intensively urbanised environment and, if present, their relative abundance in the assemblage.…”

“…We assessed the predictive power of this model by estimating the misclassification of out‐of‐bag samples (error rate) when using the model (OOB‐ER). Following the randomForest protocol suggested by Brieuc et al (2018), we first optimized the mtry parameter (number of predictors to be randomly sampled at each node in a tree). We then used the optima of each metric to run 2000 trees twice, and compared the stability of the results (correlation > 0.97 in all cases).…”

Estimating Forest Aboveground Carbon Storage in Hang-Jia-Hu Using Landsat TM/OLI Data and Random Forest Model

“…(2) number of variables to be randomly sampled at each node in a tree (Mtry), used to search for the variable that best partitions samples in the training data set and the default number is 1/3 of input variables [73]; and, (3) the minimum number of terminal nodes (Nodesize) where the default value is 5 in regression analysis [8,49].…”

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

“…GenABEL was used to perform association tests to identify SNPs potentially associated with dolphin susceptibility and resistance to Random Forest is a tree-based ensemble machine-learning tool, which is highly data adaptive, making it very useful for analysing genomic data (Chen & Ishwaran, 2012). This algorithm is particularly suited for detecting (with a high prediction accuracy) contigs that best explain variation in a response variable (Brieuc, Waters, Drinan, & Naish, 2018), and therefore loci under selection, for data sets with many thousands of SNPs and a relatively small number of samples (Chen & Ishwaran, 2012 to calculate the out-of-bag (OOB) error rate. In each RF, 125,000 trees were generated; with between two and six randomly chosen SNPs considered in each tree split (mtry; Supporting Information Table S3).…”