Covalent Docking Setup

[debanjansen48]

Issue summary

I want to execute covalent docking based virtual screening followed by ROC/enrichment / validation experiments

Steps to reproduce

<1>
I am unable to understand few things in the given protocol

gnina -r rec.pdb.gz -l conformer.sdf.gz --autobox_ligand bindingsite.sdf.gz --covalent_rec_atom A:601:FE --covalent_lig_atom_pattern '[$(n1nccc1)]' -o output.sdf.gz

A:601:FE represents A chain residue number 601,FE is iron
for my project shall I provide input A:80:SER ??
--covalent_lig_atom_pattern '[$(n1nccc1)]' hear n1nccc1 pyrazole. For my ligand shall provide the total ligand smile or the carbonyl group form covalent interaction with receptor? my PDB id is 5L2F

gnina -r 5l2f-Apo-Prep.pdb -l doripenem-prep.sdf --autobox_ligand doripenem-native.sdf --covalent_rec_atom A:80:SER --covalent_lig_atom_pattern '[$(C=O)]' -o doripenem-cov.sdf
_
(_)

---

/ | '_ \| | '_ \ / _ |
| (| | | | | | | | | (| |
_, || |||| ||_,|
/ |
|/

gnina master:617ee3b Built Dec 18 2023.
gnina is based on smina and AutoDock Vina.
Please cite appropriately.

Commandline: gnina -r 5l2f-Apo-Prep.pdb -l doripenem-prep.sdf --autobox_ligand doripenem-native.sdf --covalent_rec_atom A:80:SER --covalent_lig_atom_pattern '[$(C=O)]' -o doripenem-cov.sdf

Usage error: Could not find receptor atom A:80:SER

if i use specific atom
Commandline: gnina -r 5l2f-Apo-Prep.pdb -l doripenem-prep.sdf --autobox_ligand doripenem-native.sdf --covalent_rec_atom A:704:OG --covalent_lig_atom_pattern [$(C=O)] -o doripenem-cov.sdf

Usage error: Could not find receptor atom A:704:OG

however, in the protein file its present
ATOM 699 N SER A 80 -15.610 -18.165 43.058 1.00 14.44 N
ATOM 700 CA SER A 80 -14.463 -17.317 42.745 1.00 15.23 C
ATOM 701 C SER A 80 -14.839 -16.021 41.996 1.00 16.28 C
ATOM 702 O SER A 80 -14.027 -15.104 42.003 1.00 13.88 O
ATOM 703 CB SER A 80 -13.453 -18.115 41.900 1.00 15.79 C
ATOM 704 OG SER A 80 -12.929 -19.198 42.642 1.00 16.39 O
ATOM 705 H SER A 80 -15.675 -19.052 42.581 1.00 14.44 H
ATOM 706 HA SER A 80 -13.981 -17.022 43.680 1.00 15.23 H
ATOM 707 HB3 SER A 80 -12.609 -17.488 41.614 1.00 15.79 H
ATOM 708 HB2 SER A 80 -13.903 -18.480 40.980 1.00 15.79 H

<2>

I have total 910K ligands for setting up vs, for the each ligand how to provide smiles?

I have set of known ligands in total 11 and EC50 experimentally calculated
from DUD-E database i generated 1071 decoy structures

how to train gnina CNN using above data

Your system configuration

Operating system:LinuxMint
Compiler:gcc c++
CUDA version (if applicable):11.7
CUDNN version (if applicable):8
Python version:3

Thanks in advance. Plz assist

[dkoes]

SER is the residue name, not the atom name. The ligand pattern is a SMARTS expression for the single ligand atom you want to be covalently bonded to he specified receptor atom.

[debanjansen48]

certainly sir. SER is the residue name

it shows error if i use Atom id viz:
if i use specific atom
Commandline: gnina -r 5l2f-Apo-Prep.pdb -l doripenem-prep.sdf --autobox_ligand doripenem-native.sdf --covalent_rec_atom A:704:OG --covalent_lig_atom_pattern [$(C=O)] -o doripenem-cov.sdf

Usage error: Could not find receptor atom A:704:OG

shall i use following flag:
--covalent_rec_atom arg Receptor atom ligand is covalently bound
to. Can be specified as
chain:resnum:atom_name or as x,y,z
Cartesian coordinates

A:80:OG to define A chain SER residue 80 OG atom?

<previously used _Commandline: gnina -r 5l2f-Apo-Prep.pdb -l doripenem-prep.sdf --autobox_ligand doripenem-native.sdf --covalent_rec_atom A:80:SER --covalent_lig_atom_pattern '[$(C=O)]' -o doripenem-cov.sdf

Usage error: Could not find receptor atom A:80:SER_ >

"The ligand pattern is a SMARTS expression for the single ligand atom you want to be covalently bonded to he specified receptor atom." SMARTS of total ligand or specific part of the ligand that is the covalent warhead ?

[dkoes]

As a reminder, here is the documentation for the options:

  --covalent_rec_atom arg            Receptor atom ligand is covalently bound 
                                     to.  Can be specified as 
                                     chain:resnum:atom_name or as x,y,z 
                                     Cartesian coordinates.
  --covalent_lig_atom_pattern arg    SMARTS expression for ligand atom that 
                                     will covalently bind protein.

I suspect that 704 is not the resnum given it was 80 in your earlier attempt.

The SMARTS expression identifies a single atom. The environment of that atom can be as large or small as you'd like. Information on SMARTS can be found here: https://www.daylight.com/dayhtml/doc/theory/theory.smarts.html

[debanjansen48]

Thanks for the update..
Commandline: gnina -r 5l2f-Apo-Prep.pdb -l doripenem-prep.sdf --autobox_ligand doripenem-native.sdf --covalent_rec_atom A:80:OG --covalent_lig_atom_pattern '[CX3]=[OX1]' -o doripenem-cov.sdf
Using random seed: -626349776

0% 10 20 30 40 50 60 70 80 90 100%
|----|----|----|----|----|----|----|----|----|----|

---

mode | affinity | intramol | CNN | CNN
| (kcal/mol) | (kcal/mol) | pose score | affinity
-----+------------+------------+------------+----------
1 -10.16 0.00 0.5731 5.868
2 -10.60 0.00 0.5010 5.899
3 -8.10 0.00 0.4777 5.484
4 -11.98 0.00 0.4493 5.670
5 -6.35 0.00 0.4444 5.213
6 -10.43 0.00 0.4370 5.666
7 -5.12 0.00 0.4228 5.210
8 -4.43 0.00 0.4029 5.097
9 -6.00 0.00 0.3960 5.285
Using random seed: -626349776

0% 10 20 30 40 50 60 70 80 90 100%
|----|----|----|----|----|----|----|----|----|----|

---

mode | affinity | intramol | CNN | CNN
| (kcal/mol) | (kcal/mol) | pose score | affinity
-----+------------+------------+------------+----------
1 11.67 0.00 0.4208 5.288
2 1.55 0.00 0.4108 5.380
3 25.83 0.00 0.3709 5.058
4 33.18 0.00 0.3653 4.997
5 47.91 0.00 0.3481 5.236
6 1.12 0.00 0.3204 5.342
7 -1.23 0.00 0.3106 5.380
8 33.46 0.00 0.2967 4.734
9 4.59 0.00 0.2903 5.395

yes it works. amazing.. very easy and incredibly 1st. Awesome tool. I used various covalent docking tools. No one is as easy as 1st like gnina..

one thing I need ask you, for your kind consideration sir
like other docking tools thus gnina generate protein ligand complex for said covalent docking.