INTERGEN: a computational package applied to large-scale animal and plant genetic analysis

Background Genomic selection has become the standard genetic evaluation approach for estimating the genetic merit of animal and plant populations. That trend is mainly driven by the cost reduction for obtaining genomic data, such as single-nucleotide polymorphism (SNP) cost as novel genotyping and sequencing becomes more efficient and worldwide accessible. Although several software options are available for efficient estimation of non-genomic (i.e., traditional) genetic evaluations, the constant increase in genomic data availability requires new software that handles this data efficiently. The INTERGEN package has been constantly developing using Fortran programming language over the last two decades. The most recent of INTERGEN contains three different binaries ( intergen , intergeniod , intergenacc ). The initial developments in intergen were focused on Bayesian modeling with reaction norms, multiple-breed, and parentage uncertainty allowing for Gaussian and heavier-tailed alternatives such as Student’s t or Slash densities for variance component estimation. More recently, intergen has been constantly expanded to include the Best Linear Unbiased Prediction (BLUP), genomic module. The intergeniod and intergenacc are new binary and it intends to estimate genetic merit using iteration on data (IOD), and approximate accuracy, respectively. Results Several single-step GBLUP (ssGBLUP) single- and multi-trait models were implemented for running validations with benchmarking software ( intergen and BLUPF90). The software produced reliable estimates for both genetic merit and accuracy in all IOD and approximate accuracy analyses. Rank correlation for the genetic merit for all bivariate ssGBLUP analyses in intergeniod was equal to intergen solutions. The approximate accuracies were compared with the exact accuracies estimated in BLUPF90. The rank correlation, regression slopes, and bias in intergenacc varied based on the amount of information (i.e., genotypes and phenotypes) for each animal across the different analyses. Conclusions The INTERGEN package can be applied to the genetic analysis of large-scale datasets using Bayesian and BLUP analysis. For traditional methods such as BLUP, intergen is a flexible tool. The package allows users to adopt IOD and approximate accuracy using intergeniod and intergenacc , respectively.