Skip to main content

Table 4 Correlations between fitness values obtained by high-throughput experiments using functional complementation yeast assays [52] on 6862 GCK variants and those predicted by all the methods assessed. The best score for each metric is indicated in bold

From: FiTMuSiC: leveraging structural and (co)evolutionary data for protein fitness prediction

Method

Spearman

Pearson

FiTMuSiC

0.49

0.40

FATHMM

0.39

0.30

PROVEAN

0.41

0.33

DEOGEN2

0.51

0.35

PolyPhen-2

0.36

0.23

EVE\(^*\)

0.47

0.40

Sequence UNET

0.30

0.24

MutPred2

0.51

0.34

AlphaMissense

0.53

0.40

  1. *EVE’s predictions are available for only 6414/6862 GCK variants; missing values where set to the median
Back to article page

Human Genomics

ISSN: 1479-7364