FindSUMO: A PSSM-Based Method for Sumoylation Site Prediction

Post-translation modification by sumoylation is an important step in the regulation of many cellular processes. Existing programs predict sumoylation sites based on a consensus motif ( φ-K-X-E/D. However, ~23% of real SUMO sites do not match this motif, which makes prediction of sumoylation sites complicated. Here, we present a new method, FindSUMO, which predicts sumoylation sites using a position-specific scoring matrix. The comparison of FindSUMO with other two programs (SUMOsp and SUMOplot) indicates that FindSUMO accurately predicts sumoylation sites out of fewer total candidates and, in many cases, real sites with the highest scores. Overall, FindSUMO has at least the same performance as the other methods. While many improvements are expected, FindSUMO can also be applied to predict sumoylation sites that do not follow the consensus motif.