Hi
Thank you for sharing such an excellent short peptide design software. After using it, I have a question I’d like to ask. I used pepmlm to design some short peptides targeting my protein of interest and sorted them based on pseudo perplexity from low to high. Additionally, I performed ColabFold predictions for these short peptides with the target protein.
When comparing the prediction results (ColabFold ranked by ipTM), it seems that there is very little overlap between the top-ranked peptides in pepmlm (with low pseudo perplexity) and those with high ipTM scores in ColabFold.
I’m wondering whether ColabFold might not be well-suited for predicting short peptide complexes, or if pepmlm relies on other scoring metrics instead?
Hi
Thank you for sharing such an excellent short peptide design software. After using it, I have a question I’d like to ask. I used pepmlm to design some short peptides targeting my protein of interest and sorted them based on pseudo perplexity from low to high. Additionally, I performed ColabFold predictions for these short peptides with the target protein.
When comparing the prediction results (ColabFold ranked by ipTM), it seems that there is very little overlap between the top-ranked peptides in pepmlm (with low pseudo perplexity) and those with high ipTM scores in ColabFold.
I’m wondering whether ColabFold might not be well-suited for predicting short peptide complexes, or if pepmlm relies on other scoring metrics instead?