During recent years many protein fold recognition methods have been develop
ed, based on different algorithms and using various kinds of information. T
o examine the performance of these methods several evaluation experiments h
ave been conducted. These include blind tests in CASP/CAFASP, large scale b
enchmarks. and long-term, continuous assessment with newly solved protein s
tructures. These studies confirm the expectation that for different targets
different methods produce the best predictions, and the final prediction a
ccuracy could be improved if the available methods were combined in a perfe
ct manner. In this article a neural-network-based consensus predictor, Pcon
s, is presented that attempts this task. Pcons attempts to select the best
model out of those produced by six prediction servers, each using different
methods. Pcons translates the confidence scores reported by each server in
to uniformly scaled values corresponding to the expected accuracy of each m
odel. The translated scores as well as the similarity between models produc
ed by different servers is used in the final selection. According to the an
alysis based on two unrelated sets of newly solved proteins, Pcons outperfo
rms any single server by generating similar to8%-10% more correct predictio
ns. Furthermore. the specificity of Pcons is significantly higher than for
any individual server. From analyzing different input data to Pcons it can
be shown that the improvement is mainly attributable to measurement of the
similarity between the different models. Pcons is freely accessible for the
academic community through the protein structure-prediction metaserver at
http://bioinfo.pl/meta/.