Fig. 9
From: PRESCOTT: a population aware, epistatic, and structural model accurately predicts missense effects
Assessing ESCOTT, PRESCOTT, VARITY, gMVP and other approaches with experimental variant effect maps. A Spearman correlation coefficients are computed over 10,020 experimental variant effects for six human disease related genes on VARITY, PrimateAI, MPC and CADD on a data set proposed in [9]. B Precision-recall curves on the two human proteins PTEN and MSH2. ESCOTT is compared to gMVP and other published methods on the two datasets of pathogenic mutations, for PTEN and MSH2, proposed in [56]