HArD

Overview

HArD, HeteroAryl Descriptors database comprises DFT-computed steric and electronic descriptors of >31,500 heteroaryl substituents. The database has 65 descriptors (38 electronic, 11 steric, 16 fingerprint) computed at the M06-2X/6-31+G(d)/SMD(water)//B3LYP-D3(BJ)/6-31+G(d) level of theory.

The database can be used in three ways:

1. Website: hard.pengliugroup.com
2. Download the Excel file hard.xlsx from this repository
3. Using a python script hard.py

The last option is recommended when searching for a large list of SMILES, for example, to study correlation with reactivity/selectivity or building machine learning (ML) models.

Quick Start

1. Create a new Python virtual environment and install RDKit (2022.09.4) and pandas (1.5.3):

conda create -n hard_env python=3.11.0 -y 
conda activate hard_env
conda install -c conda-forge rdkit=2022.09.4 pandas=1.5.3 -y

2. Use curl or wget to download hard.py and hard.db:

mkdir hard && cd hard 
curl -O https://raw.githubusercontent.com/turkiAlturaifi/HArD/main/hard.py
curl -O https://raw.githubusercontent.com/turkiAlturaifi/HArD/main/hard.db

Usage

You can use the hard.py script to search for heteroaryl descriptors from SMILES strings in two ways:

1. A specific regioisomer:
   To search for a specific regioisomer, include a star in the SMILES string to represent a dummy atom. This is represented by the letter "A" in ChemDraw. For example, if you want to prepare SMILES string for 2-pyridyl, first draw a pyridine in ChemDraw, use dummy atom “A” to replace the C-H at the 2-position:

   

   Select this structure. Then go to “Edit”, “Copy As …”, “SMILES”. This would copy the SMILES string of 2-pyridyl to clipboard: [*]C1=NC=CC=C1

2. All regioisomers:
   Searching by SMILES string of the parent heteroarene would provide all matched regioisomers. For example, if you input pyridine: C1=CC=CN=C1 it will return 2-pyridyl, 3-pyridyl, and 4-pyridyl.

Command-Line Options

The hard.py script provides several options for searching and exporting data. By default, only selected descriptors are printed.

• --smi "SMILES" or -s "SMILES" : search for descriptors using a SMILES string
• --export filename.csv : export full output data to a CSV file.
• --input_file filename.txt/csv or --infile filename.txt/csv : provide a file (.txt or .csv) containing a list of SMILES (one per row). A CSV file with all the descriptors will be automatically generated.
• --full : show all descriptors
• --electronic or --elec : only electronic descriptors
• --steric : only steric descriptors
• --fingerprint : only fingerprint descriptors

Examples

Example 1: electronic descriptors of 2-pyridyl

python hard.py --smi "[*]C1=NC=CC=C1" --elec

Example 2: regioisomers of pyridine and export csv file

python hard.py --smi "C1=CC=CN=C1" --export output.csv

Example 3: List of SMILES from a file

Create a text file (e.g., smiles.txt) with each SMILES string in a separate line:

C1=CNC=C1
[*]C1=NC=CC=C1
CC1=CN=CN=C1

then run

python hard.py --input_file smiles.txt --export descriptors.csv

Generating the database:

Source codes to generate the compound library, perform high-throughput DFT calculations, and extracting the descriptors are provided in the gen_database folder.