Reader Comments

Post a new comment on this article

ambitious but ...

Posted by raik on 19 Sep 2010 at 15:31 GMT

This is a very ambitious work but it quite obviously was rushed out. My two main concerns are:
(1) about 10% of the supposed homology models are, in fact, genuine yeast structures with 100% sequence identity between template and target.
(2) presumably, many of the docking runs are "re-docking" two structures that have been taken from the same co-crystallized complex -- a situation that is very different from the real docking of "unbound" structures. The authors admit that this "may" be the reason for the surprisingly high success rate. It is really frustrating that they don't give us the actual number -- it's a simple matter of counting. Without knowing this fraction of re-docking, it is impossible to evaluate the reliability of the method or the reported success rates.

My overall impression is that of an important job, left unfinished.

No competing interests declared.