Methods for estimation of model accuracy in CASP12

被引：23

作者：

Elofsson, Arne ^{[1
,2
]}

Joo, Keehyoung ^{[3
,4
]}

Keasar, Chen ^{[5
]}

Lee, Jooyoung ^{[3
,6
]}

Maghrabi, Ali H. A. ^{[7
]}

Manavalan, Balachandran ^{[3
,6
]}

McGuffin, Liam J. ^{[7
]}

Hurtado, David Menendez ^{[1
,2
]}

Mirabello, Claudio ^{[8
]}

Pilstal, Robert ^{[8
]}

Sidi, Tomer ^{[5
]}

Uziela, Karolis ^{[1
,2
]}

Wallner, Bjorn ^{[8
]}

机构：

[1] Stockholm Univ, Dept Biochem & Biophys, Box 1031, S-17121 Solna, Sweden

[2] Stockholm Univ, Sci Life Lab, Box 1031, S-17121 Solna, Sweden

[3] Korea Inst Adv Study, Ctr Silico Prot Sci, Seoul 130722, South Korea

[4] Korea Inst Adv Study, Ctr Adv Computat, Seoul 130722, South Korea

[5] Ben Gurion Univ Negev, Dept Comp Sci, Beer Sheva, Israel

[6] Korea Inst Adv Study, Sch Computat Sci, Seoul 130722, South Korea

[7] Univ Reading, Sch Biol Sci, Reading RG6 6AS, Berks, England

[8] Linkoping Univ, Bioinformat Div, Dept Phys Chem & Biol, S-58183 Linkoping, Sweden

来源：

PROTEINS-STRUCTURE FUNCTION AND BIOINFORMATICS | 2018年 / 86卷

基金：

瑞典研究理事会; 新加坡国家研究基金会; 以色列科学基金会;

关键词：

CASP; consensus predictions; estimates of model accuracy; machine learning; protein structure prediction; quality assessment; PROTEIN-STRUCTURE PREDICTION; QUALITY ASSESSMENT; FOLD RECOGNITION; STRUCTURAL MODELS; MODFOLD SERVER; ENERGY TERMS; CONSENSUS; INTFOLD; PCONS; SCORE;

D O I：

10.1002/prot.25395

中图分类号：

Q5 [生物化学]; Q7 [分子生物学];

学科分类号：

071010 ; 081704 ;

摘要：

Methods to reliably estimate the quality of 3D models of proteins are essential drivers for the wide adoption and serious acceptance of protein structure predictions by life scientists. In this article, the most successful groups in CASP12 describe their latest methods for estimates of model accuracy (EMA). We show that pure single model accuracy estimation methods have shown clear progress since CASP11; the 3 top methods (MESHI, ProQ3, SVMQA) all perform better than the top method of CASP11 (ProQ2). Although the pure single model accuracy estimation methods outperform quasi-single (ModFOLD6 variations) and consensus methods (Pcons, ModFOLDclust2, Pcomb-domain, and Wallner) in model selection, they are still not as good as those methods in absolute model quality estimation and predictions of local quality. Finally, we show that when using contact-based model quality measures (CAD, lDDT) the single model quality methods perform relatively better.

引用

页码：361 / 373

页数：13

共 55 条

[1] Differentiable, multi-dimensional, knowledge-based energy terms for torsion angle probabilities and propensities [J].