Issue with generating `multiconformer_model2.pdb` with example 3K0N

[sai-advaith]

Hi,

I setup the github repository in my location machine (Ubuntu 22) by cloning the repository and following the instructions in the README file.

After which, I rand the code with the example files provided (qfit_protein example/qfit_protein_example/3k0n_map.mtz -l 2FOFCWT,PH2FOFCWT example/qfit_protein_example/3k0n_refine.pdb).

The code generated each residue in a separate directory A_i. However, at A_7, the code gives a warning [A/PHE7] Too many conformers generated (17100). Splitting QP scoring and the tqdm bar does not progress from there. Similar problems happen with multiple cycles.

Is there a reason why the code does not work with the examples provided on my end?

Please let me know if you need any additional details.

[stephaniewankowicz]

Hi @sai-advaith,

We apologize. There is an error with oversampling aromatic residues in the main branch. We are hoping to update the main this week, but I would suggest using the dev branch in the meantime, as that error is fixed there.

[sai-advaith]

Hi,

Thank you for your reply.

I tried setting up both, dev and main branch on my local machine a few days ago. After that, I tried to generate 5 alternate conformers in residue 113 of 3K0N's chain A using the command from README: qfit_protein qfit_protein_example/3k0n_map.mtz -l 2FOFCWT,PH2FOFCWT qfit_protein_example/3k0n_refine.pdb --residue A,113

I get similar outputs from both branches: [A/PHE113] Too many conformers generated (17100). Splitting QP scoring.

Do you know a branch and hash commit I can use to effectively generate appropriate alternate conformers using the files in the example/ directory?

Thanks!

Additional information regarding the packages:

1. python version: 3.10
2. Numpy version: 1.22
3. cvxpy version: 1.5.2
4. Pandas version: 2.2.2
5. pyscipopt version: 5.1.1
6. Scipy version: 1.14.0