Comparison of the choice of animals for re-sequencing in two maternal pig lines

Abstract

Next-generation sequencing is a promising approach for the detection of causal variants within previously identified quantitative trait loci. Because of the costs of re-sequencing experiments, this application is currently mainly restricted to subsets of animals from already genotyped populations. Imputation from a lower to a higher marker density could represent a useful complementary approach. An analysis of the literature shows that several strategies are available to select animals for re-sequencing. This study demonstrates an animal selection workflow under practical conditions. Our approach considers different data sources and limited resources such as budget and availability of sampling material. The workflow combines previously described approaches and makes use of genotype and pedigree information from a Landrace and Large White population. Genotypes were phased and haplotypes were accurately estimated with AlphaPhase. Then, AlphaSeqOpt was used to optimize selection of animals for re-sequencing, reflecting the existing diversity of haplotypes. AlphaSeqOpt and ENDOG were used to select individuals based on pedigree information and by taking into account key animals that represent the genetic diversity of the populations. After the best selection criteria were determined, a subset of 57 animals was selected for subsequent re-sequencing. In order to evaluate and assess the advantage of this procedure, imputation accuracy was assessed by setting a set of single nucleotide polymorphism (SNP) chip genotypes to missing. Accuracy values were compared to those of alternative selection scenarios and the results showed the clear benefits of a targeted selection within this practical-driven approach. Especially imputation of low-frequency markers benefits from the combined approach described here. Accuracy was increased by up to 12% compared to a randomized or exclusively haplotype-based selection of sequencing candidates.

Evaluation of the selection of key individuals for genotype imputation in Chinese yellow-feathered chicken

“…In the previous study, it was noted that there has been limited application of systematic studies utilizing the mentioned methods in chicken populations. However, consistent with prior investigations on pigs ( Dauben et al, 2022 ), based on practical data, and cattle ( Druet et al, 2014 ), based on simulated data, targeted selection techniques exhibit superior imputation performance compared to random selection. Additionally, Dauben et al reported that combining samples derived from pedigree-based and haplotype-based methodologies resulted in higher imputation accuracy than using haplotype-based samples alone, a finding that is congruent with our current study.…”

“…To identify the reference broilers, 4 distinct strategies were evaluated, as outlined in the study by Dauben et al (2022) . These approaches are designated by the letters H, P, C, and R, which respectively stand for haplotype-based, pedigree-based, combined, and random sampling.…”

“…In general, the aim of P method is to maximize the scale of the unique genomes of population. In this section, we refer to 2 pedigree-based methods mentioned by Dauben et al (2022) . To implement the P1 approach, we utilized the AlphaSeqOpt software and followed the algorithm described by Goddard et al (2011) .…”

“…While most algorithms aim to maximize the proportion of haplotypes in all animals, IWS adopts an adverse selection approach by selecting samples with a higher frequency of rare haplotypes ( Druet et al, 2014 ; Bickhart et al, 2016 ). AlphaSeqOpt is a flexible software that can run on 3 different systems, including 64-bit Linux, Mac OSX, and Windows, and has been applied for sequencing allocation in animals like pigs ( Johnsson et al, 2019 ; Dauben et al, 2022 ). AlphaSeqOpt offers 2 algorithms: one chooses key samples that cumulatively capture the maximum percentage of haplotype diversity, while the other optimizes the allocation of a fixed sequencing budget in the selected subset ( Gonen et al, 2017 ).…”

“…LPChoose is a Julia package capable of allocating sequencing resources using linear programming, which is one of the most important methods in optimization ( Cheng et al, 2021 ). The performance of various selection methods has been described in several studies that employed practical data on livestock, including cattle and pigs ( Dauben et al, 2022 ; Mcewin et al, 2022 ). However, few relevant studies have systematically investigated the selection of chickens.…”