search

dauparas / LigandMPNN

MIT License
236 stars 47 forks source link

Output an npz file with 20 mutation probabilities #2

Open PKU-Leroy opened 10 months ago

PKU-Leroy commented 10 months ago

Hi Justas, many thanks for your positive update! Can ligandMPNN output an npz file similar to ProteinMPNN to obtain the mutation probability of each site?

dauparas commented 10 months ago

Hey! It's not available at the moment, but I can easily add this functionality if it's useful.

PKU-Leroy commented 10 months ago

Yes, I am interested in testing the correlation between the predicted values from ligandMPNN and experimental values. So the mutation probabilities might be very useful. I am looking forward to seeing your code update regarding this matter.

---- Replied Message ---- | From | Justas @.> | | Date | 01/25/2024 12:43 | | To | @.> | | Cc | Ke-Wei @.>@.> | | Subject | Re: [dauparas/LigandMPNN] Output an npz file with 20 mutation probabilities (Issue #2) |

Hey! It's not available at the moment, but I can easily add this functionality if it's useful.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

PKU-Leroy commented 9 months ago

Thanks！

---- Replied Message ---- | From | @.> | | Date | 02/20/2024 12:37 | | To | @.> | | Cc | Ke-Wei @.>@.> | | Subject | Re: [dauparas/LigandMPNN] Output an npz file with 20 mutation probabilities (Issue #2) |

You can get mutation probabilities by running:

python score.py \ --model_type "ligand_mpnn" \ --seed 111 \ --single_aa_score 1\ --pdb_path "./outputs/ligandmpnn_default/backbones/1BC8_1.pdb" \ --out_folder "./outputs/single_aa_score_w_seq" \ --use_sequence 1\ --batch_size 1 \ --number_of_batches 10

This script will output a dictionary per input PDB. The value you are looking for is called out_dict["mean_of_probs"] which will have probabilities for every residue and every amino acid.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

dauparas commented 9 months ago

It should work now with:

python score.py \
        --model_type "ligand_mpnn" \
        --seed 111 \
        --single_aa_score 1\
        --pdb_path "./outputs/ligandmpnn_default/backbones/1BC8_1.pdb" \
        --out_folder "./outputs/single_aa_score_w_seq" \
        --use_sequence 1\
        --batch_size 1 \
        --number_of_batches 10
dauparas commented 9 months ago

This script will output a dictionary per input PDB. The value you are looking for is called out_dict["mean_of_probs"] which will have probabilities for every residue and every amino acid.