Alex Morehead

Failure to correctly parse all ligand chains in complex PDB file

1

Hello. It seems that with Prody version 2.4.1 (same behavior with the latest changes in `master`) not all unique ligand chains can be parsed for the attached complex (protein-ligand-nucleic acid)...

* Fixes the periodicity batching bug for `supervised_chi_loss` discovered in https://github.com/aqlaboratory/openfold/issues/381. * Fixes the shape of `unnormalized_angles_sin_cos` for (batched) angle norm loss calculation. * Also corrects the docstring for this...

Hello. Thank you all for making this work fully open-source. I had a question about the `supervised_chi_loss()` function. When constructing `chi_pi_periodic`, shouldn't the einsum equation be `...ij,jk->...ik` rather than `...ij,jk->ik`...

* This PR fixes typos and clarifies wording in the docstrings of `residue_constants.py`.

Hello. It looks to me like `batched_gather` here is an unused import. Would you all agree? https://github.com/aqlaboratory/openfold/blob/84659c93ba6f06b8a0a2646d1cf27646c003a0c6/openfold/utils/loss.py#L33

Does (or will) Foldseek support other types of molecules?

3

Hello. Thank you for making this great tool open-source. Is support for other types of molecules besides proteins on the roadmap for FoldSeek's development?

* Now overriding global `dataset_info` properties when processing CrossDock using only Ca atoms, as otherwise these values would not be updated (i.e., correct) e.g., for `amino_acid_dict` at runtime. * To...

Hi, @lucidrains. Having read the NeuralPlexer paper not too long ago, one of the things that stuck out to me about it is how vast the authors' pretraining routine for...

* Fixes a small typo in `README.md`

* Fixes `atom37_to_atom14` tensor shape errors in `all_atom_multimer.py`