Figure 1.
The mechanism and performance of iBLUP.
The figure illustrates how observed genotype with missing value is imputed by Best Linear Unbiased Prediction (BLUP). The imputation uses both relationship among markers represented by Linkage Disequilibrium (LD), and relationship among individuals represented as Identity By Decent (IBD).G and K are the genetic variance–covariance matrix and marker-based kinship matrix respectively, and the symbol represents the Kronecker product.
Table 1.
The comparison of four genotype imputation methods: iBLUP, BEAGLE, M-MM and fastPHASE.
Table 2.
Imputation accuracy characterized by minor allele*.
Table 3.
Responses of imputation accuracy on marker density and individual relationship*.
Figure 2.
(1) Blended raw data were generated from the same flow cell lane. (2) Raw data were assigned to individuals according to the barcode. (3) Assigned reads were filtered for high quality reads according to several rules, including trimming the barcode and the last low quality base etc. (4) Filtered reads were aligned with the reference sequence. (5) SNP calling and genotyping were done according to the mapping results. (6) Missing genotypes were imputed by the iBLUP algorithm.
Table 4.
Imputation accuracy of real pig sequencing data.