June 24, 2025 by Christian Loeffeld

Query Features for Substructure and Similarity Searching

This post offers a concise look at query features, a powerful and practical concept boosting substructure and similarity searching in cheminformatics. It uses a chemist’s intuition to explain how to build substructure queries that include detailed structural information. This can cover everything from stereo centers and exclude groups to variable ring sizes and atomic chain configurations, and more. This concept is implemented in the free and open-source cheminformatics platform DataWarrior.

Query features are hands-down, one of the coolest and most useful concepts in cheminformatics. The expressive power and ease of use of query features are mind blowing. Let us have a look at two simple examples to cover quite a bit of territory.

For illustration purposes only, suppose that the activity of Iptacopan is essentially triggered by the two substructures marked in red on the LHS of the image below. The corresponding query structure, a potential input for a substructure search is depicted on the RHS.

Transformation of compound structure to query structure.

The variable aspects such as chain lengths, exclude groups, and logical operators of the query are defined in the structure editor in DataWarrior as follows. Let us look at one more example in a bit more detail. In the structure editor below, you see a query structure to illustrate how to configure variable atom bridges, variable ring systems with excluded atoms and stereo centers, and lastly atomic chain composition.

Query structure configuration in structure editor.

Control panel configurations to yield the parameterization of the query structure depicted in the structure editor above. You can try out query features in Hyperspace Search for free in DataWarrior. Visit the Hyperspace site and follow the instructions on the site.

For more information on DataWarrior, check out openmolecules.org.

For related and all other inquiries, reach out to us at contact@alipheron.com.

Link to LinkedIn article.