Tram Vi, Yves Vigouroux, Philippe Cubry, Pierre Marraccini, Ha Viet Phan, Giang Ngan Khong, Valerie Poncet

Genome Biol Evol.; 2023 May 5; 15(5):evad065. doi: 10.1093/gbe/evad065.

Abstract

Humans have had a major influence on the dissemination of crops beyond their native range, thereby offering new hybridization opportunities. Characterizing admixed genomes with mosaic origins generates valuable insight into the adaptive history of crops and the impact on current varietal diversity. We applied the ELAI tool-an efficient local ancestry inference method based on a two-layer hidden Markov model to track segments of wild origin in cultivated accessions in the case of multiway admixtures. Source populations-which may actually be limited and partially admixed-must be generally specified when using such inference models. We thus developed a framework to identify local ancestry with admixed source populations. Using sequencing data for wild and cultivated Coffea canephora (commonly called Robusta), our approach was found to be highly efficient and accurate on simulated hybrids. Application of the method to assess elite Robusta varieties from Vietnam led to the identification of an accession derived from a likely backcross between two genetic groups from the Congo Basin and the western coastal region of Central Africa. Admixtures resulting from crop hybridization and diffusion could thus lead to the generation of elite high-yielding varieties. Our methods should be widely applicable to gain insight into the role of hybridization during plant and animal evolutionary history.

See https://pubmed.ncbi.nlm.nih.gov/37079743/ or

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10159586/

Figure 3: Framework for LAI of cultivated C. canephora. ELAI was performed for each individual chromosome and involved three main steps. Step 1: analyzing the genetic structure of the ancestral group, by performing sNMF on the reference set and tested set. Step 2: simulating source populations based on sNMF-estimated ancestral genotypic frequencies. Step 3: running ELAI on the tested individuals using two marker sets, that is whole-chromosome SNPs and evenly distributed SNPs. Step 4: merging the ancestry dosages inferred in the two SNP sets to determine the final consensus inference of the target chromosome.