Optimized Use of Low-Depth Genotyping-by-Sequencing for Genomic Prediction Among Multi-Parental Family Pools and Single Plants in Perennial Ryegrass (Lolium perenne L.)

Abstract

Ryegrass single plants, bi-parental family pools, and multi-parental family pools are often genotyped, based on allele-frequencies using genotyping-by-sequencing (GBS) assays. GBS assays can be performed at low-coverage depth to reduce costs. However, reducing the coverage depth leads to a higher proportion of missing data, and leads to a reduction in accuracy when identifying the allele-frequency at each locus. As a consequence of the latter, genomic relationship matrices (GRMs) will be biased. This bias in GRMs affects variance estimates and the accuracy of GBLUP for genomic prediction (GBLUP-GP). We derived equations that describe the bias from low-coverage sequencing as an effect of binomial sampling of sequence reads, and allowed for any ploidy level of the sample considered. This allowed us to combine individual and pool genotypes in one GRM, treating pool-genotypes as a polyploid genotype, equal to the total ploidy-level of the parents of the pool. Using simulated data, we verified the magnitude of the GRM bias at different coverage depths for three different kinds of ryegrass breeding material: individual genotypes from single plants, pool-genotypes from F2 families, and pool-genotypes from synthetic varieties. To better handle missing data, we also tested imputation procedures, which are suited for analyzing allele-frequency genomic data. The relative advantages of the bias-correction and the imputation of missing data were evaluated using real data. We examined a large dataset, including single plants, F2 families, and synthetic varieties genotyped in three GBS assays, each with a different coverage depth, and evaluated them for heading date, crown rust resistance, and seed yield. Cross validations were used to test the accuracy using GBLUP approaches, demonstrating the feasibility of predicting among different breeding material. Bias-corrected GRMs proved to increase predictive accuracies when compared with standard approaches to construct GRMs. Among the imputation methods we tested, the random forest method yielded the highest predictive accuracy. The combinations of these two methods resulted in a meaningful increase of predictive ability (up to 0.09). The possibility of predicting across individuals and pools provides new opportunities for improving ryegrass breeding schemes.

Genomic Prediction in Tetraploid Ryegrass Using Allele Frequencies Based on Genotyping by Sequencing

“…Ashraf et al ( 2016 ) did not correct for low accuracy of allele frequency estimates at low sequencing depth, and showed this creates a general trend of increasing genomic heritability with increasing sequencing depth. In the current study, we corrected for the effects of low accuracy at low sequencing depth, based on Cericola et al ( 2018 ), and generally see no more clear linear correlation between sequencing depth and heritability. For instance, the FILTBOTH strategy also grouped SNPs into different depth intervals, and highest estimates of genomic heritability were found for the middle to lower levels FILTBOTH2 (depth 5–10) and FILTBOTH3 (depth 10–20).…”

“…Finally, the G matrix was corrected for the extra binomial variance due to limited sequencing depth. The correction was derived by Cericola et al ( 2018 ) and simply can be done according to ploidy number and the average depth of the sample. Corrected G matrix ( G * ) was calculated by scaling down the diagonal elements of each individual as follows:…”

“…As reviewed by Poland and Rife ( 2012 ), GBS has become a flexible and low cost tool for plant genetics and breeding. It has been demonstrated that GBS can effectively generate high-density genome wide markers in a range of species (Elshire et al, 2011 ; Poland and Rife, 2012 ; Poland et al, 2012 ; Beissinger et al, 2013 ; Crossa et al, 2013 ; Zhang et al, 2015 ; Fè et al, 2016 ; Cericola et al, 2018 ). With GBS, an accurate GP model was derived for the large, complex, and polyploid wheat genome (Poland et al, 2012 ).…”

Genomic preselection with genotyping-by-sequencing increases performance of commercial oil palm hybrid crosses

“…In addition to these SNP filtering strategies, we also used several methods to compute the genomic coancestry matrices G , to check whether they could improve GS accuracies: applying corrections to G to account for the specificities of GBS (heterogeneous sequencing depth, relatively high percentage of missing data) [ 58 , 62 ], and computing G matrices from SNPs weighted according to local LD levels [ 63 , 64 ]. These results are not presented here as, with our data, they did not improve prediction accuracies.…”

Generating High Density, Low Cost Genotype Data in Soybean [Glycine max (L.) Merr.]

“…Seen frequently in plants is the use of imputation to fill missing data points in GBS data (Chung et al 2017;Chan, Hamblin, and Jannink 2016). Specially designed populations such as bi-parental, nested, and multi-parent where the founders are genotyped to a high depth and used for the reference haplotypes has been shown to boost accuracy (Bayer et al 2015;Swarts et al 2014;Tian et al 2011;Emma Huang et al 2014;Cericola et al 2018).…”