Reader Comments

Post a new comment on this article

How is the common feature determined?

Posted by dsingh on 09 Jan 2007 at 21:13 GMT

The method is based on the assumption that homologous proteins either both have or lack a protein feature
http://plosone.org/article/info:doi/10.1371/journal.pone.0000113#article1.body1.sec3.p1

"Common feature" is a somewhat nebulous term. Does it have to be some kind of sequence motif? Is the method equally reliable for different features?

RE: How is the common feature determined?

kliger replied to dsingh on 14 Jan 2007 at 08:19 GMT

Common feature could be any information about the proteins. Such information may be either experimental or computational. For example, the cellular localization of a protein, or the presence/absence of a structural motif. The feature should be conserved between homologs, and we recommend to make sure it does not affect the alignment (e.g. in this paper, we cut out the signal peptide sequences before running blastp). Theoretically, reliability should be equal for any feature; however two “technical” issues may affect reliability: (1) Is the feature always being conserved between homologs? (2) How good is the method that detects the feature?