smipoly 0.1.0

rule-based virtual polymer library generator

SMiPoly

1. What is SMiPoly?

"SMiPoly (Small Molecules into Polymers)" is rule-based virtual library generator for discovery of functional polymers. It is consist of two submodules, "monc.py" and "polg.py".
"monc.py" is a monomer classifier from a list of small molecules, and "polg.py" is a polymer repeating unit generator from the classified monomer list.

2. Current version and requirements

current version = 0.1.0
requirements

• pyhon 3.7, 3.8, 3.9, 3.10, 3.11, 3.12
• rdkit >= 2020.09.1.0 #(2019.09.3 is unavailable)
• numpy >= 1.20.2
• pandas >= 1.2.4

3. Installation and usage

3-1. Installatin

SMiPoly can be installed with pip or conda.

3-1-1. Install with pip

Create new virtual environment and activate it. To install this package, run as follows.

$pip install smipoly

3-1-2. Install with conda

Add the channel "conda-forge" if it have not been enable.

$conda config --add channels conda-forge

Create a new environment.

$conda create -n "YOUR_NEW_ENVIRONMNT_NAME" python  
or 
$conda create -n "YOUR_NEW_ENVIRONMNT_NAME" python="required version (ex. 3.10)"

Then activate it.

$conda activate "YOUR_NEW_ENVIRONMNT_NAME"

And install SMiPoly.

$conda install smipoly

Or after create and activate a new environment,

$conda install -c conda-forge smipoly

3-2. Quick start

Download 'sample_data/202207_smip_monset.csv' and 'sample_script/sample_smip_demo.ipynb' from SMiPoly repository to the same directry on your computer. Then run sample_smip_demo.ipynb. To run this demo script, Jupyter Notebook is required.

4. Module contents

4-1. monc.py

The functions of monc.py is as follows.

• extract monomers from a list of small molecules.
• classify extracted monomers into each monomer class.

The chemical structure of the small molecule compounds should be expressed in simplified molecular input line entry system (SMILES) and given as pandas DataFrame.

Functions
smip.monc.moncls(df, smiColn, minFG = 2, maxFG = 4, dsp_rsl=False)
smip.monc.olecls(df, smiColn, minFG = 1, maxFG = 4, dsp_rsl=False)

ARGUMENTS:

• df: name of the object DataFrame
• smicoln: The column label of the SMILES column, given as a str.
• minFG: minimum number of the polymerizable functional groups in the monomer for successive polymerization (default for moncls, 2: 2 or more; for olecls, 1: 1 or more)
• maxFG: maxmum nimber of the polymerizable functional groups in the monomer for successive polymerization (default 4: 4 or less)
• dsp_rsl: display classified result (default False)

Defined monomer class
By the function "moncls"

• vinylidene
• cyclic olefin
• epoxide and diepoxide
• lactone
• lactam
• hydroxy carboxylic acid
• amino acid
• cyclic carboxylic acid anhydride and bis(cyclic carboxylic acid anhydride)
• hindered phenol
• dicarboxylic acid and acid halide
• diol
• diamine and primary diamine
• diisocyanate
• bis(halo aryl)sulfone
• bis(fluoro aryl)ketone

By the function "olecls"
(The following class of compounds are also belong to the class "vinylidene" and / or "cyclic olefin".)

• acryl
• styryl
• allyl
• conjugated dienes
• vinyl ether
• vinyl ester
• maleic imide derivatives

4-2. polg.py

"polg.py" gives all synthesizable polymer repeating units starting from the classified monomer list generated by "monc.py".
For chain polymerization (polyolefins and some polyether), it gives homo and binary-copolymers. For successive (or step) polymerization, it gives homopolymer only.

Function
smip.polg.biplym(df, targ = ['all'], Pmode = 'a', dsp_rsl=False)

ARGUMENTS:

• df: name of the DataFrame of classified monomers generated by monm.moncls.
• targ: targetted polymer class. When present, it can be a list of str. The selectable elements are 'polyolefin', 'polyester', 'polyether', 'polyamide', 'polyimide', 'polyurethane', 'polyoxazolidone' and 'all' (default = ['all'])
• Pmod: generate all isomers of the polymer repeating unit ('a') or the polymer repeating unit of its representation ('r'). (default = 'a')
• dsp_rsl: display the DataFrame of the generated polymers. (default False)

Defined polymer class

• polyolefin, polycyclic olefin and their binary copolymers
• polyester (from lactone, hydroxy carboxylic acid, dicarboxylic acid + diol, diol + CO and cyclic carboxylic acid anhydride + epoxide)
• polyether (from epoxide, hindered phenol, bis(halo aryl)sulfone + diol and bis(fluoro aryl)ketone + diol)
• polyamide (from lactam, amino acid and dicarboxylic acid + diamine)
• polyimide (bis(cyclic carboxylic acid anhydride + primary diamine)
• polyurethane (diisocyanate + diol)
• polyoxazolidone (diepoxide + diisocyanate)

4-3 Sample data

The sample dataset './sample_data/202207_smip_monset.csv' includes common 1,083 monomers collected from published documents such as scientific articles, catalogues and so on.

4-4. Utilities

By using the files in './utilities' directory, one can modify or add the definition of monomers, the rules of polymerization reactions and polymer classes.
To apply the new rule(s), replace the old './smipoly/rules' directory by the new one. The files must be run according to the number assigned the head of the each filename.

• 1_MonomerDefiner.ipynb: definitions of monomers
• 2_Ps_rxnL.ipynb: rules of polymerization reactions
• 3_Ps_GenL.ipynb: definitions of polymer classes with combinations of starting monomer(s) and polymerization reaction

5. Copyright and license

Copyright (c) 2022 Mitsuru Ohno
Released under the BSD-3 license, license that can be found in the LICENSE file.

6. Publications

SMiPoly: Generation of a Synthesizable Polymer Virtual Library Using Rule-Based Polymerization Reactions
Mitsuru Ohno, Yoshihiro Hayashi, Qi Zhang, Yu Kaneko, and Ryo Yoshida
Journal of Chemical Information and Modeling 2023 63 (17), 5539-5548
DOI: 10.1021/acs.jcim.3c00329
https://doi.org/10.1021/acs.jcim.3c00329
(version 0.0.1 was used)

7. Related projects

RadonPy (Fully automated calculation for a comprehensive set of polymer properties)
https://github.com/RadonPy/RadonPy

8. Directry configuration

SMiPoly
├── src
│   └── smipoly
│       ├── __init__.py
│       ├── _version.py
│       ├── smip
│       │   ├── __init__.py
│       │   ├── funclib.py
│       │   ├── monc.py
│       │   └── polg.py
│       └── rules
│           ├── excl_lst.json
│           ├── mon_dic_inv.json
│           ├── mon_dic.json
│           ├── mon_lst.json
│           ├── mon_vals.json
│           ├── ps_class.json
│           ├── ps_gen.pkl
│           └── ps.rxn.pkl
├── LICENSE
├── pyproject.toml
├── setup.py
├── setup.cfg
├── README.md
├── sample_data
│   └── 202207_smip_monset.csv
├── sample_script
│   └── sample_smip_demo.ipynb
└── utilities
    ├── 1_MonomerDefiner.ipynb
    ├── 2_Ps_rxnL.ipynb
    ├── 3_Ps_GenL.ipynb
    └── rules/

Reference

https://future-chem.com/rdkit-chemical-rxn/
https://www.daylight.com/dayhtml_tutorials/languages/smarts/smarts_examples.html
https://www.daylight.com/dayhtml/doc/theory/theory.smarts.html