Computer-Aided Drug Design at the Durrant Lab › Forums › Gypsum-DL › Protonation problem
- This topic has 3 replies, 1 voice, and was last updated 4 years, 5 months ago by
Jacob Durrant.
- AuthorPosts
- September 23, 2020 at 2:52 pm #16059
Marcos Villarreal
GuestHello,
I am observing a inconsistent behavior in the protonation of a simple phenol molecule, running a simple example
python3 run_gypsum_dl.py –source ./examples/sample_molecules2.smi –output_folder out2 –min_ph 6 –max_ph 7 –pka_precision 1 –max_variants_per_compound 1 –separate_output_files –use_durrant_lab_filtersWhen repeating this same command I get different protonation states from time to time.
Below are the outputs when you can see that some times in the ionization step the molecules is
Mol #0: Oc1ccccc1 and after a reptetion it is ionized as Mol #0: [O-]c1ccccc1Let me know if you need more information in order to reproduce this isseue.
Best Regards, MarcosFirst time :
#——————————————————————————————————————–
Loading molecules from sample_molecules2.smi…
Desalting all molecules (i.e., keeping only largest fragment).
Contents of MolContainers
MolContainer #0 (phenol)
Mol #0: Oc1ccccc1
Ionizing all molecules…
Contents of MolContainers
MolContainer #0 (phenol)
Mol #0: Oc1ccccc1
#—————————————————————————————————————-Second time with the same command :
#—————————————————————————————————————-
Loading molecules from sample_molecules2.smi…
Desalting all molecules (i.e., keeping only largest fragment).
Contents of MolContainers
MolContainer #0 (phenol)
Mol #0: Oc1ccccc1
Ionizing all molecules…
Contents of MolContainers
MolContainer #0 (phenol)
Mol #0: [O-]c1ccccc1
#—————————————————————————————————————- - September 24, 2020 at 6:24 am #16179
Jacob Durrant
KeymasterHi Marcos. Much thanks for your interest in Gypsum. The algorithm isn’t actually deterministic, so it makes sense that you could get different outputs. Given your choice of min_ph, max_ph, and pka_precision, Gypsum is likely producing both the protonated and deprotonated form internally. But because max_variants_per_compound is 1, Gypsum must pick between the two. A certain percentage of the time it picks one, and at other times it picks the other. Hope this answer helps.
- September 27, 2020 at 3:04 am #16184
Marcos Villarreal
GuestHi Jacob,
Thank you for your answer. I realize that I did not mention that the pKa of phenol is around 9, that is why I was surprised to get the phenolate (O-) instead of the phenol (OH) at this pH range (6 to 7).Is there a combination of flags to tell gypsum to output only the most probable protonation state at a given pH?
Regards, Marcos. - September 27, 2020 at 3:10 am #16245
Jacob Durrant
KeymasterHi Marcos.You could try parameters like these:
--min_ph 7.0
--max_ph 7.0
--pka_precision 0.0
For what it’s worth, it looks like the pKa values of phenols vary quite a bit depending on other moieties: https://www.chem.ucla.edu/~cantrill/30A_S05/Acids_Bases_2.pdf
The average pKa value of the phenols used to parameterize Dimorphite-DL was particularly low: https://git.durrantlab.pitt.edu/jdurrant/dimorphite_dl/-/blob/1.2.4/site_substructures.smarts#L37
But I believe using the parameters above should always protonate your phenol molecules.
Take care,
Jacob
- AuthorPosts
- The topic ‘Protonation problem’ is closed to new replies.