Table 3.

Phenotypic variance in LT50 explained (R ²) by significant markers identified in GWAS.

A
Model	R 2 with kinship	R 2 without kinship	Number of loci
SNP-based	0.30	0.48	8
Haplotype-based	0.26	0.45	71
B
Marker	R 2 with kinship	R 2 without kinship
SNP-based model
LG1_10733810	0.05	0.09
LG1_11608488	0.05	0.19
scaf_836_20601	0.05	0.10
scaf_1414_9867	0.04	0.01
LG1_20019082	0.02	0.08
scaf_153_142327	0.02	0.00
LG6_12095124	0.01	0.00
LG3_13453807	0.01	0.12
Haplotype-based model
LG1:11608425-11608558_31,2	0.05	0.19
LG1:10733751-10733981_31	0.05	0.09
LG2:26124549-26124674_01	0.04	0.15
LG1:4339801-4340029_24	0.04	0.03
LG3:13143867-13144034_06	0.04	0.12
scaf_670:26744-26998_19	0.03	0.08
scaf_28637:389-568_03	0.02	0.03

Regression analysis was performed, using allele frequencies of significant markers as predictors of LT50. The R ² is given for both SNP- and haplotype-based models that either account for or do not account for kinship among accessions. A, R ² for the regression models including all markers simultaneously; B, R ² for the regression models including single markers.

¹One haplotype had a p-value below the Bonferroni threshold of significance in the GWAS but it was included in further analyses because it was very close to being significant and included one significant SNP. ²Numbers at the end of each haplotype correspond to the variant number within HTP loci.