helixonprotein / binding-ddg-predictor Goto Github PK

It seems that ΔΔG by different methods could have different meaning. In GeoPPI, the Github page stated that a positive ΔΔG means mutation increase binding affinity. Is that the same with ΔΔG calculated by binding-ddg-predictor? Please clarify. Thanks.

Hi, I followed the instructions and installed the package. However, it seemed to be in an endless loop when I ran the example WT and mutant pdb provided. After troubleshooting, I narrowed down my problem to cudatoolkit might not be compatible with my GPU. I wanted to ask if there is a GPU requirement to run your ddg-predictor?

during binding-ddg-predictior, I got the below error. model.pt file exists in the data directory.

The paper mentions that "The structures of mutated complexes were sampled using the Rosetta relax".
We followed the above procedure to generate the structures after single point mutation and tested them using 'model.pt'. The Pearson correlation coefficient of our retest was only 0.29, while the Pearson correlation coefficient of the S1131 dataset in the paper was 0.65, which is very different from it.
It would help us a lot if you could provide us with the structure of the mutation. Or can you provide us with the parameters of your rosetta relax.

Hi, which one of the following formulae use in binding-ddg-predictor for ddg calculation?
1- ddg = dg/mutant - dg/wild
or
2- ddg = dg/wild - dg/mutant

Could share the training, test, and validation splits which you used on SKEMPI v2?

Also, can you highlight how the example mutation PDB file that is included in the github repo was constructed? It'd be great if you could share these details such as the Rosetta Cartesian DDG method or if there are other packages that we could use as it would help us understand the method better and re-use it.

Looking forward to your reply.

Aritra

Thank you for this impressive contribution! Could you add a license file when you have a moment?

Best wishes,

-Cyrus

Dear authors,
If I want to use another antigen (PEG .RNA DNA..), how can I train the data (Antibody-Bind (AB-Bind) database) to build my model.pt ?
thanks
Yeng-Tseng

Hi, I have read your research paper " Deep learning guided optimization of human antibody against SARS-CoV-2 variants with broad neutralization" and have a question that where is the Mutation Libraries from?

Where the description in citation are:

• Variant-specific amino acid substitutions in the RBD were first created to build complexes with different antigen variants. For each antigen variant, single-point mutations in CDR of the antibody were enumerated.
• While the initial structure of P36-5D2 bound with the RBD, the neutralizing activity of P36-5D2 against SARS-CoV-2 WT, Beta, Gamma, and Delta was then utilized to analyze and select CDR mutations. Based on those data, our method generated an in silico mutation library of antibody CDRs

So what are the details?
Thank you.