Identification and analysis of genetic variants obtained from the whole-genome sequencing of Nellore bulls

Abstract

Strategies to use the complete DNA sequence in animal genetic evaluations of several species have been developed due to advances in next-generation sequencing techniques associated with decreasing in sequencing costs and due to an expected increasing of prediction accuracies by the possibility of using causal mutations. Aligner and variant calling methods are of paramount importance to use sequencing data in genetic analyses. There are differences among methods regarding the number and phase concordance of part of the identified genotypes. These differences can influence the results of analysis performed to genetic evaluations which are directly dependent of variant discovery. Thus, the use of an inadequate method could compromise the results of all subsequent analyses or, at least, to reduce the expected accuracy gain. In the present project, there will be used data from the complete DNA sequence and genotyping by the panel Illumina Bovine HD chip of 150 Nellore bulls selected by their importance for the breed. There will be compared two aligners (BWA e HISAT2) and two variant calling methods (SAMtools e GATK) in a factorial scheme. In this way, there will be tested four combinations to the process of identification of genetic variants (SNPs and indels) in the population (BWA-SAMtools, BWA-GATK, HISAT2-SAMtools e HISAT2-GATK). The platforms will be compared based on the Non-Reference Sensitivity (NRS) and Non-Reference Discrepancy (NRD) statistics. The NRS measures the methodology sensibility degree in which a value of NRS equal to unit means a perfect agreement between the sequencing and genotyping results, while the NRD statistics denotes the false positive rate. With the definition of the better strategy for variant identification, there will be performed the variant annotation as well as the genome linkage disequilibrium characterization. There will be also investigated the imputation accuracy of different SNP panel densities to complete genome sequence with a subsequent imputation of a database with more than 10,000 animals. The developing of this project should contribute to the obtaining and utilization process of genomic data generated from the whole-genome sequencing of beef cattle. (AU)