Open MarcAmil30 opened 1 year ago
cyangNYU
commented
1 year ago Hi, Could you check the Lin_F9 (in pK) value? If Lin_F9 (in pK) is a positive value but XGB you obtained a negative score, it would be some issues for your feature calculation. To be specific, could you share the protein-ligand complexes that you have used? The typical range one should expect for score (in pK) is (0, 14)
MarcAmil30
commented
1 year ago Thank you for the reply.
Sure, I have attached a link below to a subset of the dataset. The subset includes the same ligand already docked at different sites of one protein (therefore only needs scoring).
The link: https://drive.google.com/drive/folders/1c89ygcqMkpn4NUzYZRyGgwlTA73ofP4o?usp=sharing (Please let me know if you can access it)
cyangNYU
commented
1 year ago Hi, I found the reason for negative XGB value (in pK): the ligand structures have steric clashes with the protein structure. Actually, I don't think these ligand sdf files are docked poses (see below figure), but delta_LinF9_XGB is for re-scoring the docked poses or native crystal poses. So, what you need to do is based on docked poses to calculate delta_LinF9_XGB score.
You can based on the sites to re-dock these ligand structures using Smina Suite (https://github.com/cyangNYU/Lin_F9_test), and then perform delta_LinF9_XGB calculation.
BTW, I have tried the SITEA calculation based on docked pose:
python script/runXGB.py test02/modified_original_protein.pdb test02/SITEA_docked.pdb | grep 'XGB'
XGB (in pK) : 6.001
Hello :)
When I attempt to follow the instructions and score my complex after docking, I encounter negative results, such as XGB (in pK) with a value of -50. I would like to kindly ask about the potential reasons for obtaining this negative score. Furthermore, I'm curious to know if a negative score is considered normal, and if not, what is the typical range one should expect for both the score and the output.
Thanks