Accurate prediction of DnaK-peptide binding via molecular modelling and experimental data

Molecular chaperones are essential elements of the protein quality control machinery that govern translocation and folding of nascent polypeptides, refolding and degradation of misfolded proteins, and activation of a wide range of client proteins.

The prokaryotic heat-shock protein DnaK is the E. coli representative of the ubiquitous Hsp70 family, which specializes in the binding of exposed hydrophobic regions in unfolded polypeptides. Accurate prediction of DnaK binding sites in E. coli proteins is an essential prerequisite to understanding the precise function of this chaperone and the properties of its substrate proteins.

In order to map DnaK binding sites in protein sequences, we have developed an algorithm that combines sequence information from peptide binding experiments and structural parameters from homology modelling.

We show that this combination significantly outperforms either single approach. The final predictor had a Matthews correlation coefficient (MCC) of 0.819 when assessed over the 144 tested peptide sequences to detect true positives and true negatives.

To test the robustness of the learning set, we have conducted a simulated cross-validation, where we omit sequences from the learning sets and calculate the rate of repredicting them. This resulted in a surprisingly good MCC of 0.703.

The algorithm was also able to perform equally well on a blind test set of binders and non-binders, of which there was no prior knowledge in the learning sets.

http://www.ncbi.nlm.nih.gov/pubmed/19696878

Please use this reference to cite Limbo:

Van Durme J, Maurer-Stroh S, Gallardo R, Wilkinson H, Rousseau F, Schymkowitz J. Accurate prediction of DnaK-peptide binding via homology modelling and experimental data. PLoS Comput Biol. 2009 Aug;5(8):e1000475.

Options Guide

Best overall prediction

This option is selected by default and is recommended for the most accurate chaperone binding site prediction.

For DnaK binding, this prediction setting has 99% specificity and 77.2% sensitivity. This implies that for our benchmark peptide set, this predictor setting predicts 77.2% of the true positives with a concurrent amount of 1% false positives. In practice, every heptapeptide that scores above the value of 11.08, is a predicted DnaK binder with this prediction setting.

High specificity prediction

This option aims for high specificity (less false positives) but could be at a cost of finding less true positives.

For DnaK binding, this prediction setting has 100% specificity and 70.5% sensitivity. This implies that for our benchmark peptide set, this predictor setting predicts 70.5% of the true positives with no false positives. In practice, every heptapeptide that scores above the value of 12.5, is a predicted DnaK binder with this prediction setting.

High sensitivity prediction

This option aims for high sensitivity (more true binders) but at a cost of less specificity (more false positives).

For DnaK binding, this prediction setting has 91% specificity and 93.2% sensitivity. This implies that for our benchmark peptide set, this predictor setting predicts 93.2% of the true positives but with a concurrent amount of 9% false positives. In practice, every heptapeptide that scores above the value of 8.26, is a predicted DnaK binder with this prediction setting.

Logically, this setting should detect more predicted binders than the above high specificity setting.

Best per sequence

This option prints as output the best scoring peptide per input sequence.

For DnaK binding, the best scoring heptapeptide per input sequence is printed to the output screen.

Custom

With this option users can specify a custom specificity value. E.g. a value of 80 will allow 20% false positives in the prediction. When this value is set to zero it allows every possible peptide motif to be scored and printed to the screen.

For DnaK binding the peptide motif is a heptapeptide.