Table 1.

Correlations between yield and selected regions

Yield	N	RSNP-A	PSNP-A	RSNP-B	PSNP-B	RSNP-C	PSNP-C	RRegion	PRegion
Wuhan, 2011	289	0.085	0.15	0.26	5.5 × 10−6	−0.020	0.74	0.26	9.1 × 10−6
Wuhan, 2012	295	0.21	2.0 × 10−4	0.20	7.3 × 10−4	0.17	4.1 × 10−3	0.26	5.0 × 10−6
Lingshui, 2012	283	0.19	1.6 × 10−3	0.24	6.0 × 10−5	0.13	3.1 × 10−2	0.24	4.2 × 10−5
Normalized average	295	0.23	5.1 × 10−5	0.34	1.3 × 10−9	0.12	3.4 × 10−2	0.38	2.7 × 10−11
BLUP breeding value	295	0.32	1.9 × 10−8	0.53	5.6 × 10−23	0.15	1.1 × 10−2	0.58	6.8 × 10−28

To obtain the normalized average for the yield data of three field experiments, we subtracted the mean value for each field experiment and averaged when multiple observations were available. BLUP breeding values for yield were obtained using ridge regression BLUP (49) from yield data of three field experiments. R_SNP-A and P_SNP-A were calculated based on 446,593 SNPs with an allele frequency difference (ΔDAF) greater than 0.3 between IndI and IndII; R_SNP-B and P_SNP-B were calculated based on 74,724 SNPs of SNP-A located in selected regions; R_SNP-C and P_SNP-C were calculated based on 312,497 SNPs of SNP-A departed from 100 kb of selected regions; and R_Region and P_Region were calculated using the number of selected regions in each accession containing derived haplotypes. N, number of phenotyped accessions; P, P value of Spearman's rank correlation test; R, Spearman’s rank correlation coefficient.