evaluation scrpit

[pengzhangzhi]

Hi there,

First of all, thank you soo much for opening this code! I need some help with running the inference on other pdb samples. In your readme, you provide the code to infer from the examples, while I have some other cases to test. Are there any instructions on how to run the inference only given a pdb id like (7BSD_A) or a pdb structure (7BSD_A.pdb)? An easy and straightforward example would be much helpful!

Best,
Zhangzhi

[YidongSong]

Thank you for your interest in our research. You can of course test your own new examples. Here we give detailed instructions on how to run the model only given a pdb structure.
1、Follow the readme document to prepare prottrans feature. You can also refer to the guide (https://github.com/agemagician/ProtTrans) if you have any questions.
2、Predict the protein structures using ESMFold. If you have any doubts, please refer to this document (https://github.com/facebookresearch/esm).
3、Calculate the dssp features with reference to https://onlinelibrary.wiley.com/doi/abs/10.1002/bip.360221211.
4、Change the path to the data, and you can run the model for your own cases.

[pengzhangzhi]

Thanks! do you have a script that can do all those steps in once? how do you evaluate on other proteins?

[YidongSong]

Thank you for your interest in our research. We have provided a complete pipeline in predict.py. You can only provide the fasta file and the model will automatically complete the prediction. This discussion is divided into two situations. If the required features are already available, the existing features will be used. Otherwise, the required features will be automatically calculated, and the corresponding prediction will be made. Thanks for your attention, I hope the answer is helpful to you.

[pengzhangzhi]

Hi yidong, thanks! Would you also provide an instruction on how to set up and run it with only sequence input?