Skip to main content
. 2021 Jan 21;12:510. doi: 10.1038/s41467-020-20847-0

Fig. 3. Distribution of predicted scores of de novo missense variants by MVP and other methods.

Fig. 3

For each method, we normalized all predictions by rank percentile, and used two-sided Mann–Whitney U test to assess the statistical significance of the difference between cases and controls. CHD: congenital heart disease; ASD: autism spectrum disorder; controls: unaffected siblings from the ASD study. Number of de novo missense variants compared: CHD: 1486; ASD: 2050; controls: 838.