Introduction

Common Groups is a project for molecular structure-based classification of chemical substances into groups with known environmental health hazards. This is part of a broader effort to develop information infrastructure for chemical hazard assessment and alternatives assessment.

This documentation is specifically about commongroups, a software package written in Python. It provides computational methods to identify sets of chemicals belonging to structurally-defined classes.

Purpose

Certain groups or classes of chemical substances are of interest for their environmental or toxicological hazard characteristics. Chemical hazard screening methods and regulatory agencies reference several hundred such groups. However, determining the actual set of substances that might belong to each group is usually left up to someone else.

The Common Groups project aims to address this gap though the collaborative development of molecular structure-based “definitions” for chemical groups of interest, supported by software tools and shared information resources. The collaborative effort will be convened and documented by the Chemical Hazard Data Commons.

The goal of the Common Groups software tool is to apply basic cheminformatics techniques to:

• Find all substances within a larger set that belong to a given group.
• Classify individual substances into the correct group(s).
• Perform these functions automatically for a large number of groups all at once.

We imagine the possible applications of this project to include answering questions such as: What dithiocarbamates are in this list of compounds? or Does this new compound belong to any chemical groups associated with endocrine disruption?

Frequently asked questions

Does this program use structural similarity searching? No, it uses substructure matching. The idea is that the groups of chemicals we want to identify are defined by precise criteria, rather than inferred by similarity. We see this approach as complementary to similarity searching methods, with each approach having different advantages.

Is this a tool for identifying toxicophores? No. Toxicophore identification is part of the rational basis for defining groups of substances by structure, and is therefore a background condition, not a function, of this project.

Isn’t this project limited by the current state of knowledge linking individual groups of chemicals to individual hazard endpoints? Yes. The purpose of this project is not to create new knowledge about compound groups, but to identify potentially new associations between specific compounds and hazards, based on existing knowledge.