Analysis of substructural variation in families of enzymatic proteins with applications to protein function prediction.

The decrease in false positive matches, resulting from use of a motif ensemble, occurred because true positive matches tended to match multiple consensus motifs within the ensemble with low LRMSD, while many false positive matches have only marginally significant LRMSD to a single consensus motif, and applying multiple testing correction to the final set of matches for a given false positive often caused a single marginally significant match to move outside of the significance threshold.

also in 5,076 other papers