Fig. 2 | Nature Communications

Fig. 2

From: Detecting heritable phenotypes without a model using fast permutation testing for heritability and set-tests

Fig. 2

Discrepancy in p-values in a cis-eQTL study. p-values from 10,000 permutations, compared to GCTA p-values assuming asymptotics (in log scale). Evaluated on 22,171 gene expression profiles in whole-blood samples, from the GTEx dataset, with 338 individuals. Sites with \(\hat h^2 = 0\) (8604 profiles) omitted for clarity of presentation, with 99.995% CIs. Parametric p-values are often smaller than the exact p-values obtained by the permutation test, frequently by several orders of magnitude, resulting in many false positives

Back to article page