0 avis
Comparison of High-Throughput Genotyping Technologies in Maize: Increase in number of regions detected in association genetics
Archive ouverte
Edité par CCSD -
Recent progress in genotyping and resequencing technologies can improve our understanding of the mechanisms involved in quantitative trait variation by performing genome-wide association studies on large diversity panels. Circa 250 Dent inbred lines with narrow phenology variation were selected within the DROPS project. Hybrids with a flint tester were evaluated for male flowering time and plant heigth at seven sites in Europe, during two years (2012 and 2013), and for two treatments (watered and drought). We genotyped this panel using 50K Infinium HD Illumina array (~50,000 SNP Ganal et al., 2011), 600K Axiom Affymetrix array (~600,000 SNP Unterseer et al., 2014) and Genotyping By Sequencing approach developed by Cornell university (~600,000 SNP; Elshire et al., 2011; Glaubitz et al., 2014). We studied the impact of increasing the marker density on (i) the estimates of kinship and population structure, (ii) the accuracy of the QTL positions, and (iii) the gain in power (number of significant markers identified). We also addressed the issues of the genotyping reproducibility and the imputation quality for GBS, considering the 50K Illumina array as the reference. Genotyping are highly reproducible (c. 100% for 50K vs 600K and c. 99% for 50K vs GBS) and the imputation quality for the GBS were high (c. 97%). GBS lead to a much higher proportion of rare alleles compared to Illumina and Affymetrix arrays. Increased marker density, using 600K and GBS, leads to the detection of new regions with marker-trait associations and a higher precision of QTL positions. However, each technology leads to detection of associations in specific regions, and thus none of these three technologies seems able to capture all of the genomic regions involved in trait variation. The suggestion is that the density and/or distribution of SNPs genotyped by GBS and 600K is still not sufficient to identify all relevant genomic regions.
Consulter en ligne
Suggestions
Du même auteur
