There are currently eight prediction methods either online or in development/testing. Each one performs the same task — predicting functional associations between pairs of proteins (currently human-only) — using the same query and result syntax. However, each predictor differs in the sources of evidence it considers and the computational and/or statistical methods it uses to make predictions based on that data.
CODA (hosted at UCL): evolutionary relatedness based on domains found fused together in other species
engineDB (hosted at ITB): detection of functionally analogous proteins via GO annotations
GOSS (hosted at UCL): semantic similarity of GO annotations, by GO subtree (see below)
GECO (hosted at UCL): correlated patterns of gene expression from microarray experiments
hiPPI (hosted at UCL): homology-based inheritance of protein-protein interactions from public databases
iHOP (hosted at CNIO): inference of gene/protein relationships by text mining
JACOP (hosted at SIB): unsupervised clustering and classification based on detection of homologous sub-sequences
PIPS (hosted at Dundee): Bayesian prediction of protein-protein interactions using various sources of evidence
SpindleP (hosted at CBS): prediction of mitotic spindle proteins using artificial neural networks
CODA is available in two versions, one using CATH domains and one using Pfam domains.
iHOP is also available in two versions, the ‘classic’ version which has more reliable scores but slightly lower coverage, and the iHOP-FuncNetConn service which has better coverage at the expense of score reliability.
GOSS is available in three versions, one each for the three Gene Ontology subtrees: biological process (BP), cellular component (CC) and molecular function (MF). The GOSS algorithm is quite similar to engineDB. We recommend using engineDB if you are looking for overall similarity of annotations and do not wish to make a distinction between the different subtrees, but using GOSS if you care specifically about one or two kinds of annotation but not all three.
All of the FuncNet predictors can be queried directly, too, by SOAP-compliant tools. See their WSDL files for reference.
Juan Ranea and Ian Morilla at the University of Malaga designed the score integration process. Our EMBRACE liaison is Erik Bongcam-Rudloff at Uppsala University. We are grateful for the technical assistance of ENFIN’s Florian Reisinger, and would like to thank Michael Thomas Flanagan whose Java scientific library we used.
