A Kmer-based paired-end read (KPR) de novo assembler and genotyper to genotype major histocompatibility complex class I (MHC-I) alleles for the dog

The major histocompatibility complex class I (MHC-I) genes are highly polymorphic among individuals. MHC-I genotyping is required for determining the antigen-binding specificity of each MHC-I molecule in an individual. Numerous tools have been developed for human MHC-I genotyping using deep sequencing data such as RNA-seq; however they do not work for the dog, due to very limited information for canine alleles. To address this issue, we developed a Kmer-based paired-end read (KPR) de novo assembler and genotyper, which first assemble paired-end RNA-seq reads mapped to the MHC-I regions into contigs de novo and then genotype each contig. Our KPR tools are validated by Sanger sequencing, simulation and published genotype data. Applying our KPR tools on the published RNA-seq data of 158 tumor and 64 normal samples from 158 dogs, we have achieved a genotyping success rate of 86%, which includes 133 tumor and 57 normal samples from 142 dogs. We have identified 39 known alleles and 83 new alleles of high confidence in these dogs, yielding a more comprehensive MHC-I allele diversity landscape for the dog.