Imputed WGS data of 626 birds

Genotype imputation was performed with two-step approach from 55 to 600 K, and then imputed to WGS. Before Genotype imputation, pre-phasing was executed in Beagle 4.1 with default parameter (Browning and Browning, 2016). Firstly, using 450 birds with 600 K chip data as a reference panel, these 194 birds were imputed from 55 K to 600 K chip data using Beagle 4.0 with pedigree. And then merge 600 K chip data of these 194 birds and 450 birds using VCFtools. Secondly, all of 644 birds with 600 K chip data were imputed to WGS data using a combined reference panel Beagle 4.1 with default parameter. The combined reference panels included 24 key individuals from the yellow-feather dwarf broiler population and 311 birds with WGS data from diverse chicken breeds. More detail information could be found in our previous study (Ye et al., 2019). After performed genotype imputation, quality control of the imputed WGS data was conducted using PLINK (Purcell et al., 2007) with the criteria of SNP call rate > 95%, individual call rate > 97%, MAF > 0.5%, and Hardy-Weinberg equilibrium P-value > 1.0e-6. In addition, individuals would be excluded who existed Mendelian errors. Finally, the remaining 626 individuals and 11,173,020 SNPs were used for further analysis.