RZooRoH: An R package to characterize individual genomic autozygosity and identify homozygous-by-descent segments

Abstract

Identifying homozygous-by-descent (HBD) regions in individual genomes is highly valuable to infer the recent history of populations and to provide insights into trait architecture. Here, we present the RZooRoH R-package that implements an efficient and accurate model-based approach to identify HBD segments. The underlying hidden Markov model partitions the genome-wide individual autozygosity into different age-related HBD classes while accounting for genotyping errors and genetic map information. The RZooRoH package is user-friendly and versatile, accepting either genotyping or sequencing (including low-coverage) data in various formats. Through numerical maximization and parallelization, computational performances were improved compared to our initial Fortran implementation of the model. The package allows to evaluate and compare various models defined by their number of HBD classes and it also provides several graphical functions that help interpretation of the results. RZooRoH is an efficient tool that proves particularly suited for sub-optimal datasets (e.g. low marker density, individual low-coverage sequencing, uneven marker spacing) and for individuals from populations with complex demographic histories. RZooRoH is available from CRAN: https://CRAN.R-project.org/package=RZooRoH.