Chemical Curation

This repository provides a Python script for parsing and curation of chemical data. The curation process follows the procedure recommended in the following paper:

https://pubs.acs.org/doi/10.1021/ci100176x

Eventually, this repository will include the procedures recommended in the following:

https://pubs.acs.org/doi/abs/10.1021/acs.jcim.6b00129

Curation process

1. Data aggregation

• sdf and csv files are supported, some basic searching for desired fields is implemented

2. Removal of mixtures, inorganics, and salts

• currently using MolVS Standardizer

3. Structure standardization

• currently using MolVS Standardizer

4. Duplicate averaging/removal

• InChI keys are used to determine if structures are the same
• Users can provide the threshold for determining if activities for the same molecule are close enough to average, default threshold is one log unit

5. Manual review

• Any issues throughout the process are stored in a text file for manual review