Table 4

Accuracy of imputation in samples from African Americans

Accuracy


Missing

Random AA (258)

AA close to YRI (223)

AA close to CEU (35)


Complete missing

Overall

85.66%

(0.26, 0.75, 0.93, 1)

87.00%

(0.27, 0.77, 0.94, 1)

97.14%

(0.11, 0.91, 1.00, 1)


0.95 P.P.

96.70%

(0.00, 0.92, 0.99, 1)

96.97%

(0.00, 0.93, 0.99, 1)

100.00%

(0.00, 0.97, 1.00, 1)


80% missing

Overall

85.92%

(0.29, 0.75, 0.93, 1)

87.08%

(0.26, 0.77, 0.94, 1)

96.43%

(0.07, 0.89, 1.00, 1)


0.95 P.P.

96.79%

(0.00, 0.92, 0.99, 1)

97.06%

(0.00, 0.93, 0.99, 1)

100.00%

(0.00, 1.00, 1.00, 1)


Impact of splitting samples from African Americans into groups based on their similarity to the Yorubans and Caucasians on the imputation accuracy. The 1st column reports imputation accuracy when YRI haplotypes are used on the whole set, the 2nd column reports imputation accuracy when YRI haplotypes are used on a cluster of subjects close to YRIs, the 3rd column reports imputation accuracy when CEU haplotypes are used on a cluster of subjects close to CEUs.

Zhao et al. BMC Genetics 2008 9:85   doi:10.1186/1471-2156-9-85

Open Data