Example 3: Build a Combinatorial Library¶

This notebook briefly introduces the molli combine workflow which allows to substitute attachment points on the molecules in a combinatorial fashion.

Hardware Specification for Rerun¶

Desktop workstation with 2x (AMD EPYC 7702 64-Core) with total of 128 physical and 256 logical cores, 1024 GB DDR4 with Ubuntu 22.04 LTS operating system.

[1]:

import molli as ml
ml.visual.configure()

Overview.¶

We start with the file phosphorus_core.mol2, in which we define the three attached pseudoatoms labeled AP1, AP2 and AP3. These are the “attachment points”: the pseudoatoms that define the direction of the substitution. These atoms need to be replaced with the groups found in another file: substituents.cdxml. We will be generating two libraries: achiral and chiral phosphines.

[2]:

mol = ml.load("phosphorus_core.mol2")
for a in mol.atoms:
    print(a)
mol

Atom(element=P, isotope=None, label='P', formal_charge=0, formal_spin=0)
Atom(element=Unknown, isotope=None, label='AP1', formal_charge=0, formal_spin=0)
Atom(element=Unknown, isotope=None, label='AP2', formal_charge=0, formal_spin=0)
Atom(element=Unknown, isotope=None, label='AP3', formal_charge=0, formal_spin=0)

You appear to be running in JupyterLab (or JavaScript failed to load for some other reason). You need to install the 3dmol extension:
jupyter labextension install jupyterlab_3dmol

[2]:

Molecule(name='P', formula='P1 Unknown3')

Import prerequisite molecules as molli collections¶

[3]:

!molli compile phosphorus_core.mol2 -o P.mlib --overwrite

Matched 1 files for importing.
Importing molecules: 100%|██████████████████████| 1/1 [00:00<00:00, 2411.91it/s]

[4]:

!molli parse substituents.cdxml -o substituents.mlib --hadd --overwrite

Parsing substituents.cdxml: 100%|██████████████| 21/21 [00:00<00:00, 497.99it/s]

Achiral phosphine ligands¶

The objective here is to create a library of phosphines in which the P atom is not stereogenic, which means that at least two substituents need to be identical. We are going to achieve this by using molli’s command line interface.

Create a disubstituted phosphine library with `--mode same`¶

This ensures that two substituents will be identical

[5]:

!molli combine --help

usage: molli combine [-h] -s <substituents.mlib>
                     [-m {same,permutns,combns,combns_repl}]
                     [-a ATTACHMENT_POINTS] [-n 1] [-b 1] -o <combined.mlib>
                     [-sep SEPARATOR] [--hadd]
                     [--obopt [ff maxiter tol disp ...]] [--overwrite]
                     cores

Combines two lists of molecules together

positional arguments:
  cores                 Base library file to combine wth substituents

options:
  -h, --help            show this help message and exit
  -s <substituents.mlib>, --substituents <substituents.mlib>
                        Substituents to add at each attachment of a core file
  -m {same,permutns,combns,combns_repl}, --mode {same,permutns,combns,combns_repl}
                        Method for combining substituents
  -a ATTACHMENT_POINTS, --attachment_points ATTACHMENT_POINTS
                        Label used to find attachment points
  -n 1, --nprocs 1      Number of processes to be used in parallel
  -b 1, --batchsize 1   Number of molecules to be processed at a time on a
                        single core
  -o <combined.mlib>, --output <combined.mlib>
                        File to be written to
  -sep SEPARATOR, --separator SEPARATOR
                        Name separator
  --hadd                Add implicit hydrogen atoms wherever possible.
  --obopt [ff maxiter tol disp ...]
                        Perform openbabel optimization on the fly. This
                        accepts up to 4 arguments. Arg 1: the forcefield
                        (uff/mmff94/gaff/ghemical). Arg 2: is the max number
                        of steps (default=500). Arg 3: energy convergence
                        criterion (default=1e-4) Arg 4: geometry displacement
                        (default=False) but values ~0.01-0.1 can help escape
                        planarity.
  --overwrite           Overwrite the target files if they exist (default is
                        false)

[6]:

!molli combine P.mlib -s substituents.mlib -a AP1 -a AP2 -m same -o R2P.mlib --hadd --overwrite

Will create a library of size 21
100%|█████████████████████████████████████████████| 1/1 [00:00<00:00, 18.76it/s]

Finally, we can attach the last remaining substituent to finish the achiral phosphine library.

[7]:

!molli combine R2P.mlib -a AP3 -s substituents.mlib --hadd --obopt UFF 1000 1e-4 0.02 --overwrite -n1 -b8 -o R3P_achiral.mlib

Will create a library of size 441
100%|███████████████████████████████████████████| 63/63 [02:50<00:00,  2.70s/it]

In order to make the calculation faster, some of molli workflows (eventually, all of them) implement parallelization. See the acceleration for yourself!

[8]:

!molli combine R2P.mlib -a AP3 -s substituents.mlib --hadd --obopt UFF 1000 1e-4 0.02 --overwrite -n8 -b8 -o R3P_achiral.mlib

Will create a library of size 441
100%|███████████████████████████████████████████| 63/63 [00:25<00:00,  2.49it/s]

Chiral phosphine ligands¶

The objective here is to create a library of phosphines in which the P atom is stereogenic, which means that all substituents need to be different. We are going to achieve this by using molli’s command line interface.

The magic here is in the use of -m combns parameter of the molli combine workflow. This means that for each

[9]:

!molli combine P.mlib -s substituents.mlib -a AP1 -a AP2 -a AP3 -m combns -o R3P_chiral.mlib --hadd --obopt UFF 1000 1e-4 0.02 --overwrite -n1 -b8

Will create a library of size 1330
100%|█████████████████████████████████████████| 167/167 [08:35<00:00,  3.09s/it]

Jupyter also allows for special commands, so the command %mlib_view allows for direct visualization of molecules in a MoleculeLibrary without needing to run a full command to open it. The syntax is as follows:

%mlib_view <LIB_PATH> <KEY>

[10]:

%mlib_view R3P_chiral.mlib P_15_17_14