Closed ahof1704 closed 5 months ago
Thanks for the question, you can just use the log likelihood from ProClassConditioner
for the content probability. It is a trained model for conditioning.
But you just use any DSSP programs (e.g. https://github.com/ShintaroMinami/PyDSSP ) to compute secondary structure contents, which will be more accurate.
closing this issue for now. feel free to reopen or submit a new issue if you have follow-up questions.
Hi,
In figures 5g and 5h of your paper, you indicate that Chroma can predict the alpha and beta content of the generated molecules. Could you indicate the code you use for this prediction?
Thanks Antonio