Bayesian hybrid index and genomic cline estimation with the R package gghybrid

Area of Science:

• Evolutionary Biology
• Genetics
• Bioinformatics

Background:

• Admixture, the interbreeding of individuals from differentiated populations, is a widespread phenomenon with significant implications for adaptive evolution, reproductive isolation, and speciation.
• Genomic cline analysis is a powerful method for inferring gene flow patterns and identifying loci under selection in admixed populations, especially when geographic clines are unsuitable.
• Analyzing admixture proportions and genomic clines is crucial for understanding evolutionary processes in natural hybrid zones.

Purpose of the Study:

• To introduce gghybrid, a novel R package designed for Bayesian estimation of hybrid indices and genomic clines.
• To provide a flexible and user-friendly tool for analyzing admixture at the genomic level across various ploidy levels.
• To facilitate the study of evolutionary processes such as adaptive introgression and reproductive isolation in admixed populations.

Main Methods:

• Development of the gghybrid R package for Bayesian estimation of hybrid indices and Fitzpatrick's logit-logistic genomic clines.
• Utilizes bi-allelic data and is compatible with Structure input file format, supporting any ploidy level.
• Incorporates functions for independent analysis of individuals (hybrid index) and loci (cline estimation), enabling parallelization.

Main Results:

• gghybrid offers simple and flexible Bayesian estimation of hybrid indices and genomic clines.
• The package is suitable for both small and large datasets and can handle any ploidy.
• Simulations indicate that using the widely applicable information criterion (waic) is preferred for model comparison, and Bayesian posterior distributions/p-values should be avoided for locus selection.

Conclusions:

• gghybrid is a valuable new tool for researchers studying admixture and its evolutionary consequences.
• The package enhances the ability to infer gene flow and identify loci under selection in admixed populations.
• Best practices for model selection and locus identification in genomic cline analysis are highlighted, emphasizing the utility of waic over traditional p-values.