MC:pred is program that predicts the missed cleavage of tryptic bonds within protein sequences. It is based on a Support Vector Machine (SVM) trained on observed peptides in Peptide Atlas (S. cerevisiae, C. elegans and D. melanogaster). The output from the SVM approach is a score from 0-1, with the missed cleavages resulting in a higher score. A binary threshold of 0.5 can be applied to distinguish missed from cleaved sites. Increasing the threshold will increase the precision (PPV) of the missed cleavage prediction, but will sacrifice the recall (sensitivity). Predictions are provided for tryptic peptides of length >5.

In addition, the previous information theory approach for predicting missed cleavages (Siepen et al., 2007) is provided.

To use the program select a fasta file or paste in your sequence and select the prediction type.

The prediction types available are:
SVM - Predicts missed cleavages based on the SVM trained data
Info Theory - Predicts missed cleavages based on the previous information theory approach by Siepen et al.

If you have any problems or queries, please email craig.lawless@manchester.ac.uk.