Potential of Low‐Coverage Genotyping‐by‐Sequencing and Imputation for Cost‐Effective Genomic Selection in Biparental Segregating Populations

Abstract

Genotyping‐by‐sequencing (GBS) is an alternative genotyping method to single‐nucleotide polymorphism (SNP) arrays that has received considerable attention in the plant breeding community. In this study we use simulation to quantify the potential of low‐coverage GBS and imputation for cost‐effective genomic selection in biparental segregating populations. The simulations comprised a range of scenarios where SNP array or GBS data were used to train the genomic selection model, to predict breeding values, or both. The GBS data were generated with sequencing coverages (x) from 4x to 0.01x. The data were used either nonimputed or imputed by the AlphaImpute program. The size of the training and prediction sets was either held fixed or was increased by reducing sequencing coverage per individual. The results show that nonimputed 1x GBS data provided comparable prediction accuracy and bias, and for the used measurement of return on investment, outperformed the SNP array data. Imputation allowed for further reduction in sequencing coverage, to as low as 0.1x with 10,000 markers or 0.01x with 100,000 markers. The results suggest that using such data in biparental families gave up to 5.63 times higher return on investment than using the SNP array data. Reduction of sequencing coverage per individual and imputation can be leveraged to genotype larger training sets to increase prediction accuracy and larger prediction sets to increase selection intensity, which both allow for higher response to selection and higher return on investment.

Parent-progeny imputation from pooled samples for cost-efficient genotyping in plant breeding

“…Thus, whereas the marker concordance rate is a direct measure of 390 imputation accuracy, the GEBV concordance can be understood as measuring it 391 indirectly through its effects on WGS. Other studies investigating the use of imputed 392 marker scores for WGS used the correlation between predicted and true genetic values 393 (commonly referred to as the "prediction accuracy") as indirect measures of imputation 394 uncertainty [32,35]. We decided against this, however, because the prediction accuracy 395 depends on many other factors that are independent of the genotyping and imputation 396 process, such as the trait heritability or genetic architecture [48].…”

“…Many applications in statistical genetics, including estimation of whole genome 603 marker effects and calculation of GEBV, do not require hard genotype calls and accept 604 fractional scores proportional to the posterior probability. Carrying over the uncertainty 605 around each marker score into the subsequent analysis, as done in this study, weights 606 each score by the chance of it being incorrect and thus acts as a buffer against 607 imputation error [35]. Indeed, the average certainty of correct calls was always 608 considerably higher than that of incorrect calls (Figure 4).…”

“…Obtaining genotypes from DNA sequence data, termed genotyping by sequencing 36 (GBS) [17] emerged as another approach to reduce genotyping costs and increase scale. 37 This approach efficiently generates high volumes of genotypic data and holds particular 38 promise for applications in plant breeding and genetics [33][34][35]. Because GBS methods 39 typically result in a large amount of missing data [36], genotype imputation is an 40 integral component of this technology [33,35,37].…”

“…The step-by-step procedure was as follows 292 1. Sequenceability of each marker (seq k ) was sampled from a Gamma distribution 293 with shape and rate of 4 [35]. 294 2.…”

“…At minimum two reads could accurately 442 capture the allele dosage at such a locus, but under Binomial sampling two reads will 443 still fail to detect heterogeneity 50% of the time. Application of low-coverage GBS to 444 heterozygous or heterogeneous material therefore requires explicit accounting for read 445 sampling uncertainty[35].…”