Mycorrhiza: genotype assignment using phylogenetic networks

Abstract

Motivation The genotype assignment problem consists of predicting, from the genotype of an individual, which of a known set of populations it originated from. The problem arises in a variety of contexts, including wildlife forensics, invasive species detection and biodiversity monitoring. Existing approaches perform well under ideal conditions but are sensitive to a variety of common violations of the assumptions they rely on. Results In this article, we introduce Mycorrhiza, a machine learning approach for the genotype assignment problem. Our algorithm makes use of phylogenetic networks to engineer features that encode the evolutionary relationships among samples. Those features are then used as input to a Random Forests classifier. The classification accuracy was assessed on multiple published empirical SNP, microsatellite or consensus sequence datasets with wide ranges of size, geographical distribution and population structure and on simulated datasets. It compared favorably against widely used assessment tests or mixture analysis methods such as STRUCTURE and Admixture, and against another machine-learning based approach using principal component analysis for dimensionality reduction. Mycorrhiza yields particularly significant gains on datasets with a large average fixation index (FST) or deviation from the Hardy-Weinberg equilibrium. Moreover, the phylogenetic network approach estimates mixture proportions with good accuracy. Availability and implementation Mycorrhiza is released as an easy to use open-source python package at github.com/jgeofil/mycorrhiza. Supplementary information Supplementary data are available at Bioinformatics online.

Genome‐scale phylogeography resolves the native population structure of the Asian longhorned beetle, Anoplophora glabripennis (Motschulsky)

“…To identify markers for downstream diagnostic tool development, we wished to identify SNPs that could accurately assign individuals to their source populations. To do so, we used Mycorrhiza 0.0.28, a genotype assignment software that employs machine learning and phylogenetic networks to identify informative SNPs that can assign individuals to their source population (Georges‐Filteau et al, 2020 ). The SNPs were ranked by discriminatory power based on mutual information.…”

Coping with environmental constraints: Geographically divergent adaptive evolution and germination plasticity in the transcontinentalPopulus tremuloides

“…For an overall picture of the population structure, we performed principal component analyses (PCAs) with plink v1.90b5.3 (Purcell et al, 2007) and visualized the results in R using ggplot2. To assess detailed structure, we used the program Mycorrhiza 0.0.28, a machine learning approach that utilizes phylogenetic networks to identify features that encode evolutionary relationships among samples that are then supplied to a Random Forest classifier to identify structure (Georges‐Filteau et al, 2020). The program requires pre‐identified groups as input to assess the accuracy of samples assigned to a group.…”

Genomic biosurveillance of forest invasive alien enemies: A story written in code

“…Already genome‐derived tools have been designed for the identification of tree rust fungi and of Phytophthora species and lineages, and this is being applied in eradication programs targeting the EU1 lineage of P. ramorum in Oregon (Bergeron et al, ; Feau et al, , ; Leboldus et al, ). Novel efficient and rapid algorithms using machine‐learning can provide a probability of assignment given the genomic reference library and associated metadata developed for each invasive species (Georges‐Filteau, Hamelin, & Blanchette, ). The outcome of the genomic predictions can then be integrated into a decision support system to provide a user with easily interpretable outcomes.…”