Imperial College London


Faculty of Natural SciencesDepartment of Life Sciences

Director Centre for Bioinformatics



+44 (0)20 7594 5212m.sternberg Website




306Sir Ernst Chain BuildingSouth Kensington Campus






BibTex format

author = {Jiang, Y and Oron, TR and Clark, WT and Bankapur, AR and D'Andrea, D and Lepore, R and Funk, CS and Kahanda, I and Verspoor, KM and Ben-Hur, A and Koo, DCE and Penfold-Brown, D and Shasha, D and Youngs, N and Bonneau, R and Lin, A and Sahraeian, SME and Martelli, PL and Profiti, G and Casadio, R and Cao, R and Zhong, Z and Cheng, J and Altenhoff, A and Skunca, N and Dessimoz, C and Dogan, T and Hakala, K and Kaewphan, S and Mehryary, F and Salakoski, T and Ginter, F and Fang, H and Smithers, B and Oates, M and Gough, J and Toronen, P and Koskinen, P and Holm, L and Chen, C-T and Hsu, W-L and Bryson, K and Cozzetto, D and Minneci, F and Jones, DT and Chapman, S and Dukka, BKC and Khan, IK and Kihara, D and Ofer, D and Rappoport, N and Stern, A and Cibrian-Uhalte, E and Denny, P and Foulger, RE and Hieta, R and Legge, D and Lovering, RC and Magrane, M and Melidoni, AN and Mutowo-Meullenet, P and Pichler, K and Shypitsyna, A and Li, B and Zakeri, P and ElShal, S and Tranchevent, L-C and D},
doi = {10.1186/s13059-016-1037-6},
journal = {Genome Biology},
title = {An expanded evaluation of protein function prediction methods shows an improvement in accuracy},
url = {},
volume = {17},
year = {2016}

RIS format (EndNote, RefMan)

AB - BackgroundA major bottleneck in our understanding of the molecular underpinnings of life is the assignment of function to proteins. While molecular experiments provide the most reliable annotation of proteins, their relatively low throughput and restricted purview have led to an increasing role for computational function prediction. However, assessing methods for protein function prediction and tracking progress in the field remain challenging.ResultsWe conducted the second critical assessment of functional annotation (CAFA), a timed challenge to assess computational methods that automatically assign protein function. We evaluated 126 methods from 56 research groups for their ability to predict biological functions using Gene Ontology and gene-disease associations using Human Phenotype Ontology on a set of 3681 proteins from 18 species. CAFA2 featured expanded analysis compared with CAFA1, with regards to data set size, variety, and assessment metrics. To review progress in the field, the analysis compared the best methods from CAFA1 to those of CAFA2.ConclusionsThe top-performing methods in CAFA2 outperformed those from CAFA1. This increased accuracy can be attributed to a combination of the growing number of experimental annotations and improved methods for function prediction. The assessment also revealed that the definition of top-performing algorithms is ontology specific, that different performance metrics can be used to probe the nature of accurate predictions, and the relative diversity of predictions in the biological process and human phenotype ontologies. While there was methodological improvement between CAFA1 and CAFA2, the interpretation of results and usefulness of individual methods remain context-dependent.
AU - Jiang,Y
AU - Oron,TR
AU - Clark,WT
AU - Bankapur,AR
AU - D'Andrea,D
AU - Lepore,R
AU - Funk,CS
AU - Kahanda,I
AU - Verspoor,KM
AU - Ben-Hur,A
AU - Koo,DCE
AU - Penfold-Brown,D
AU - Shasha,D
AU - Youngs,N
AU - Bonneau,R
AU - Lin,A
AU - Sahraeian,SME
AU - Martelli,PL
AU - Profiti,G
AU - Casadio,R
AU - Cao,R
AU - Zhong,Z
AU - Cheng,J
AU - Altenhoff,A
AU - Skunca,N
AU - Dessimoz,C
AU - Dogan,T
AU - Hakala,K
AU - Kaewphan,S
AU - Mehryary,F
AU - Salakoski,T
AU - Ginter,F
AU - Fang,H
AU - Smithers,B
AU - Oates,M
AU - Gough,J
AU - Toronen,P
AU - Koskinen,P
AU - Holm,L
AU - Chen,C-T
AU - Hsu,W-L
AU - Bryson,K
AU - Cozzetto,D
AU - Minneci,F
AU - Jones,DT
AU - Chapman,S
AU - Dukka,BKC
AU - Khan,IK
AU - Kihara,D
AU - Ofer,D
AU - Rappoport,N
AU - Stern,A
AU - Cibrian-Uhalte,E
AU - Denny,P
AU - Foulger,RE
AU - Hieta,R
AU - Legge,D
AU - Lovering,RC
AU - Magrane,M
AU - Melidoni,AN
AU - Mutowo-Meullenet,P
AU - Pichler,K
AU - Shypitsyna,A
AU - Li,B
AU - Zakeri,P
AU - ElShal,S
AU - Tranchevent,L-C
AU - Das,S
AU - Dawson,NL
AU - Lee,D
AU - Lees,JG
AU - Sillitoe,I
AU - Bhat,P
AU - Nepusz,T
AU - Romero,AE
AU - Sasidharan,R
AU - Yang,H
AU - Paccanaro,A
AU - Gillis,J
AU - Sedeno-Cortes,AE
AU - Pavlidis,P
AU - Feng,S
AU - Cejuela,JM
AU - Goldberg,T
AU - Hamp,T
AU - Richter,L
AU - Salamov,A
AU - Gabaldon,T
AU - Marcet-Houben,M
AU - Supek,F
AU - Gong,Q
AU - Ning,W
AU - Zhou,Y
AU - Tian,W
AU - Falda,M
AU - Fontana,P
AU - Lavezzo,E
AU - Toppo,S
AU - Ferrari,C
AU - Giollo,M
AU - Piovesan,D
AU - Tosatto,SCE
AU - del,Pozo A
AU - Fernandez,JM
AU - Maietta,P
AU - Valencia,A
AU - Tress,ML
AU - Benso,A
AU - Di,Carlo S
AU - Politano,G
AU - Savino,A
AU - Rehman,HU
AU - Re,M
AU - Mesiti,M
AU - Valentini,G
AU - Bargsten,JW
AU - van,Dijk ADJ
AU - Gemovic,B
AU - Glisic,S
AU - Perovic,V
AU - Veljkovic,V
AU - Veljkovic,N
AU - Almeida-e-Silva,DC
AU - Vencio,RZN
AU - Sharan,M
AU - Vogel,J
AU - Kansakar,L
AU - Zhang,S
AU - Vucetic,S
AU - Wang,Z
AU - Sternberg,MJE
AU - Wass,MN
AU - Huntley,RP
AU - Martin,MJ
AU - O'Donovan,C
AU - Robinson,PN
AU - Moreau,Y
AU - Tramontano,A
AU - Babbitt,PC
AU - Brenner,SE
AU - Linial,M
AU - Orengo,CA
AU - Rost,B
AU - Greene,CS
AU - Mooney,SD
AU - Friedberg,I
AU - Radivojac,P
DO - 10.1186/s13059-016-1037-6
PY - 2016///
SN - 1474-760X
TI - An expanded evaluation of protein function prediction methods shows an improvement in accuracy
T2 - Genome Biology
UR -
UR -
UR -
VL - 17
ER -