Optimised amino acid specific weighting factors for unbound protein docking
Cologne University Bioinformatics Center (CUBIC), University of Cologne, Zuelpicher Str. 47, 50674 Koeln, Germany
BMC Bioinformatics 2006, 7:344 doi:10.1186/1471-2105-7-344Published: 14 July 2006
One of the most challenging aspects of protein-protein docking is the inclusion of flexibility into the docking procedure. We developed a postfilter where the grid-representation of proteins for docking is extended by an optimised weighting factor for each amino acid.
For up to 86% of the evaluated complexes a near-native structure was within the top 5% of the ranked prediction output. The weighting factors obtained by the optimisation procedure correlate to a certain extent with the flexibility of the amino acids, their hydrophobicity and with their propensity to be in the interface.
Use of the optimised amino acid specific parameters yields a strong increase of near-native structures on the first ranks of the prediction.