question from a non-bioinformatics student - (Jul/26/2006 )
I used two different on-line programs for the prediction of SUMOylation (protein modification) target motifs in any given protein sequence. The sites are:
-SUMOplot (http://www.abgent.com/doc/sumoplot): I think it just analyzes sequences related to the consensus motif
-SUMOsp (http:/bioinformatics.lcd-ustc.org/sumosp/): they use two strategies (GPS and MotifX, ??????) to determine the putative modification target motifs
I obtained a completely different result with each one!! And the targets obtained with the second one were more divergent from the consensus motif. Could anybody explain me (in simple words, my background on bioinformatics is not too wide) why?
Thanks in advance, and sorry cause maybe my question is too simple
I haven't checked the links however, if the predictors are from different labs or built using different datasets then it is perfectly reasonable for the predictions to be different - chances are they will have not been built in the same way...
welcome to the painful world of feature/structure prediction.