Compare affinity scores for different proteins

[MaximePolicarpo]

Hi !

I was wondering if one could compare the affinity scores computed with different proteins ?

I have a set of 50 bitter taste receptors. I want to score their affinity to different bitter compounds (for example quinine, dextromethorphan, diphenhydramine ...). I then want to cluster the receptors based on their affinity scores to those different compounds. Would such an analysis make sense ?

I was also wondering if one could provide a heterodimer as a PDB files, or if the program currently only works with one protein ?

Thanks for any help provided !

Maxime

[patjiang]

Hello Maxime,

I think it would be pretty feasible to compare the affinity scores! From the paper (which is trained on the PDBBind2020 dataset, the values correspond to micromolar). I think that kind of analysis would make sense, but it would be far more interesting to also look at where the small molecules bind to the protein of interest.

I have used a heterodimer in the past, and it totally works

Good hunting!

Patrick Jiang

[MaximePolicarpo]

Hello Patrick,

Thank you a lot for your answer ! I will try to make this analysis and see if the results make sense, also regarding the true ligands known for those receptors. I can keep this issue open and update it once it will be done, it might be useful for people wanting to do something similar !

Regarding the heterodimer, would you have a PDB file to sent me ? I tried to run DynamicBind with a taste receptor heterodimer (see the attached PDB file), but did not seem to work :/ (otherwise, for bitter taste receptors alone, it worked perfectly fine)

Here is the command I ran :
python run_single_protein_inference.py fold_t1r1_t1r3_chicken_model_0.pdb T1R_ligand_hts.csv --hts --savings_per_complex 3 --inference_steps 30 --header Chicken_T1R --python miniconda3/envs/dynamicbind/bin/python --miniconda3/envs/relax/bin/python

fold_t1r1_t1r3_chicken_model_0.pdb.zip

Thanks for your help and guidance :D

All the best,

Maxime