Skip to main content
. 2009 May;182(1):343–353. doi: 10.1534/genetics.108.100289

TABLE 4.

Loss of accuracy of GEBVs with genotyping of selected low-density (SLD) SNPs as a percentage of the accuracy obtained by high-density BayesB in generations after training and with 100, 500, and 1000 simulated QTL

No. of simulated QTL Generation
Method 2 3 4
SLD-BB-110 1.1 (±3.2) 0.8 (±3.2) 0.7 (±4.2)
100 SLD-BB-40 3.1 (±3.1) 2.0 (±3.2) 1.7 (±4.2)
SLD-FSS-0.01 9.0 (±2.4) 8.1 (±3.3) 8.7 (±4.0)
SLD-FSS-110 18.2 (±3.2) 23.1 (±3.2) 24.9 (±4.2)
SLD-BB-110 2.5 (±3.2) 1.9 (±4.8) 1.5 (±5.4)
500 SLD-BB-40 7.8 (±3.2) 6.5 (±4.8) 4.02 (±5.0)
SLD-FSS-0.01 10.7 (±2.7) 10.6 (±4.0) 12.1 (±4.1)
SLD-FSS-110 12.5 (±2.8) 14.2 (±2.8) 18.8 (±4.3)
SLD-BB-110 3.0 (±3.3) 2.6 (±3.8) 2.3 (±4.9)
1000 SLD-BB-40 9.2 (±3.3) 9.7 (±3.7) 8.4 (±4.9)
SLD-FSS-0.01 12.8 (±3.2) 17.7 (±4.1) 21.0 (±4.6)
SLD-FSS-110 13.7 (±3.5) 18.4 (±3.8) 22.0 (±4.2)

One thousand individuals were used to estimate HD-SNP effects (±SE, based on 25 replicates). SLD-BB-110 and SLD-BB-40 refer to selection of 110 and 40 SNPs using BayesB; SLD-FSS-0.01 and SLD-FSS-110 refer to selection of SNPs by forward stepwise selection with a significance level of 0.01 or with a total of 110 SNPs selected.