Figure 3

Cumulative distribution plots for the six deleterious prediction scores. The X-axis represents the prediction scores, ordered by the deleteriousness. The lowest and the highest scores for each prediction algorithm indicate the neutral and damaging nsSNP, respectively. For each prediction program, the score threshold for defining damaging SNPs is indicated by a vertical dashed green line (threshold for LRT is at 0.999). Five sets of SNP data are shown in each plot. Black solid lines: data from homozygous nsSNPs of four African individuals; blue solid lines: data from homozygous nsSNPs of eight Caucasian individuals; red dashed lines: data from all disease-causing nsSNPs (n = 24703 nsSNPs in 1789 proteins); black dashed lines: data from disease-causing nsSNPs found in homozygous nsSNP-containing genes of the four African individuals (n = 348 nsSNPs in 23 genes); blue dashed lines: data from disease-causing nsSNPs found in homozygous nsSNP-containing genes of the eight Caucasian individuals (n = 547 nsSNPs in 44 genes). All disease-causing nsSNPs were retrieved from MSV3D [14] and SwissVar [15].