Table 1 CASP14 performances.

From: Estimation of model accuracy by a unique set of features and tree-based regressor

A

 

Group name

MCC(50)

MESHI_consensus

0.746

MESHI

0.742

DAVIS-EMAconsensus

0.728

ModFOLDclust2

0.724

MUfoldQA_G

0.723

EMAP_CHAE

0.707

UOSHAN

0.696

Yang_TBM

0.692

B

 

Group name

RMSE

DAVIS-EMAconsensus

0.0673

MUfoldQA_G

0.0723

MESHI_consensus

0.0724

MESHI

0.0725

ModFOLDclust2

0.0735

EMAP_CHAE

0.0739

Yang_TBM

0.0804

UOSHAN

0.0836

C

 

Group name

LOSS

MULTICOM-CONSTRUCT

0.0735

MULTICOM-AI

0.0792

MESHI

0.0793

MULTICOM-CLUSTER

0.0802

MUfoldQA_G

0.082

MESHI_consensus

0.084

BAKER-ROSETTASERVER

0.084

BAKER-experimental

0.0845

  1. The tables present the top-scoring groups31,71,72,73,74,75 by three measures: MCC(50) (A), RMSE (B), and LOSS (C). Results are reproduced from the CASP website at (https://predictioncenter.org/casp14). Note that in the CASP site RMSE and LOSS are referred to as ”differences (predicted vs observed)” and ”difference from the best”, respectively, and their performances are depicted as percentages. The servers ”Seder2020” and ”Seder2020hard” that submitted an EMA prediction for a single target were omitted.
Back to article page