Skip to main content
. 2010 Oct 5;26(22):2867–2873. doi: 10.1093/bioinformatics/btq559

Table 2.

Computation time of two software implementations to estimate kinship coefficients in three sets of GWAS SNP data

Summary of genome scan data
Computing time
Index No. of SNPs No. of samples No. of pairs KING PLINK
1 3 079 857 269 36 046 2 min 2 h 9 min
2 324 748 602 180 901 1 min 1 h 13 min
3 549 338 2 454 3 009 832 25 min 28 h 30 min

The computation time refers to the time to estimate kinship coefficients for all pairs of individuals, excluding overhead costs such as the time to load data into the computer memory. The two KING implementations (the robust algorithm and the algorithm assuming homogeneous samples) took a similar amount of computational time. This computation time can be estimated reliably as the analysis time for the entire data minus the analysis time for only the within-family data. The unit of computation time is in minutes hours. All computation was performed on and Intel Xeon with 3.20 GHz processor.