WHISCY prediction

 

How does WHISCY work?

WHISCY requires a protein structure and a sequence alignment. First, it identifies a master sequence, the sequence that best matches the structure.

The sequence distance (amount of mutation) between the master sequence and all sequences is estimated. This determines the amount of expected mutation.

Then, for each residue, the expected mutation is compared with the observed mutation. Less change than expected means conservation, translated into a positive WHISCY score.

Next, the interface propensity is taken into account.

Phenylalanines, for example, are likely to be in a protein-protein interface, so all phenylalanines receive a higher score. Lysines are much less likely to be in a protein-protein interface, so lysines receive a lower score.

 

Finally, all scores are smoothed over the surface of the protein structure.

Interfaces often form patches, so that neighbours of interface residues often are interface residues, too. The smoothing means that the scores of these neighbours are taken into account.