Detection of pig genome regions determining production traits using an information theory approach

The objectives of the study were to detect the SNPs associated with pig production traits and estimate interactions between the most informative pairs of SNPs using entropy analysis. 1296 pigs were genotyped and phenotyped. The following traits were recorded: backfat, days to 110 kg, feed conversion ratio, percent lean meat and terminal line index for Maxgro™. After validation of the genotypic data, 50 951 SNPs were considered for further analysis. Entropy and conditional entropy for each SNP were estimated. Mutual information for most informative SNPs was performed to estimate an interaction between pairs of SNPs. Generally, the majority of loci showed relatively small contributions to the traits. The largest number of SNPs with high conditional entropy coefficients was identified on chromosome 1. The most important SNPs and pairs of them for all traits were mainly located on chromosomes: 1, 4, 6, 7, 9, 14, 16 and 17. High mutual information was estimated for SNPs located nearby. Many of the most prominent SNPs are localised within candidate genes affecting pig production traits.