<p>Identifying hybrid individuals is essential for conservation management of threatened species and delimiting species boundaries. The MONGRAIL&#xa0;2.0 software package applies Bayesian inference to classify individuals as purebred or hybrid/backcrossed using genome sequences of putative hybrids and pure individuals from two reference populations. The program evaluates six genealogical classes spanning two generations: purebreds from each population, F1 hybrids, F2 hybrids, and backcrosses to each parental population. MONGRAIL&#xa0;2.0 uses a posterior predictive distribution that accounts for uncertainty in population haplotype frequencies, and correctly marginalizes over haplotypes while modeling linkage and recombination. It can use the exact likelihood for linked variants or highly accurate approximations that increase the number of SNPs that can be simultaneously analyzed per chromosome. The software accepts either VCF or custom genotype formats and outputs posterior probabilities for each genealogical class. MONGRAIL&#xa0;2.0 is implemented in C for computational efficiency and is freely available at <a href="https://github.com/mongrail/mongrail2">https://github.com/mongrail/mongrail2</a>.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Mongrail 2.0: Bayesian inference of hybrids using population genomic data

  • Sneha Chakraborty,
  • Bruce Rannala

摘要

Identifying hybrid individuals is essential for conservation management of threatened species and delimiting species boundaries. The MONGRAIL 2.0 software package applies Bayesian inference to classify individuals as purebred or hybrid/backcrossed using genome sequences of putative hybrids and pure individuals from two reference populations. The program evaluates six genealogical classes spanning two generations: purebreds from each population, F1 hybrids, F2 hybrids, and backcrosses to each parental population. MONGRAIL 2.0 uses a posterior predictive distribution that accounts for uncertainty in population haplotype frequencies, and correctly marginalizes over haplotypes while modeling linkage and recombination. It can use the exact likelihood for linked variants or highly accurate approximations that increase the number of SNPs that can be simultaneously analyzed per chromosome. The software accepts either VCF or custom genotype formats and outputs posterior probabilities for each genealogical class. MONGRAIL 2.0 is implemented in C for computational efficiency and is freely available at https://github.com/mongrail/mongrail2.