Skip to main content

Table 6 Comparing errors in estimation of allele frequencies by filtering off uncalled SNPs.

From: Validation of pooled genotyping on the Affymetrix 500 k and SNP6.0 genotyping platforms using the polynomial-based probe-specific correction

NoCall cutoff SNPs Analysed Correlation MAD error 95% CI
- 500568 (100%) 0.989 0.031 (0.0360-0.0362)
45 500488 (99.98%) 0.989 0.031 (0.0360-0.0362)
40 500353 (99.96%) 0.989 0.031 (0.0359-0.0362)
35 500036 (99.89%) 0.989 0.031 (0.0359-0.0361)
30 499222 (99.73%) 0.989 0.031 (0.0358-0.0360)
25 497469 (99.38%) 0.989 0.030 (0.0356-0.0358)
20 493792 (98.65%) 0.989 0.030 (0.0354-0.0356)
15 485709 (97.03%) 0.989 0.030 (0.0351-0.0353)
10 466758 (93.25%) 0.990 0.030 (0.0346-0.0348)
5 415562 (83.02%) 0.990 0.030 (0.0334-0.0337)
0 198749 (39.7%) 0.993 0.028 (0.0288-0.0291)
  1. Actual allele frequencies obtained from individual genotyping were compared with estimated allele frequencies from pooled genotyping at various "NoCall" cutoffs.