The way I use DNA motif (PWMs) features in the prediction is to do motif scanning using Fimo. And I used a p-value threshold to create a binary value. However, I think a continuous value would be better. Of course, I can use p-value as a continuous value. The thing is that Fimo is running very slow if you set p-value threshold close to 1. Can anyone provide me a motif scanning tool that is capable to output some scores other than p-value or q-value? This might refer to Energy-base model. In a general sense, my question is that how do you use DNA motifs in the prediction? Thanks!