Table 1 Fitness prediction results of the benchmarked methods on the 5811 variants used in the CAGI6 HMBS challenge [27]. The best score for each metric is indicated in bold
From: FiTMuSiC: leveraging structural and (co)evolutionary data for protein fitness prediction
Method | CAGI6 | Kendall | Spearman | Pearson | RMSD |
|---|---|---|---|---|---|
FiTMuSiC | \(\checkmark\) | 0.30 | 0.43 | 0.42 | 0.39 |
\(\mathcal {SNP}\) | \(\checkmark\) | 0.27 | 0.39 | 0.38 | 0.43 |
\(\mathcal {POP}\) | \(\checkmark\) | 0.15 | 0.22 | 0.24 | 0.44 |
ELAPSIC team | \(\checkmark\) | 0.30 | 0.42 | 0.43 | 0.43 |
CalVEIR team | \(\checkmark\) | 0.31 | 0.45 | 0.36 | 0.51 |
FATHMM |  | 0.16 | 0.23 | 0.17 | – |
PROVEAN |  | 0.21 | 0.31 | 0.30 | – |
DEOGEN2 |  | 0.22 | 0.32 | 0.20 | – |
PolyPhen-2 |  | 0.21 | 0.28 | 0.22 | – |
EVE\(^{\textbf {*}}\) |  | 0.29 | 0.42 | 0.43 | – |
Sequence UNET |  | 0.21 | 0.30 | 0.30 | – |
MutPred2 |  | 0.25 | 0.37 | 0.34 | – |
AlphaMissense |  | 0.32 | 0.46 | 0.41 | – |