SimFinder

SimFinder is a graph-based similarity searching tool for performing a topological search of a chemical compound database. It is based on the C-Tree technology.

Features and Capabilities

• Maintains connectivity information of topological fragments present in a compound. Hence, improved quality of results compared to traditional representation techniques such a fingerprints.
• Scales to millions of compounds.

Method

The process of building the index structure from graphs obtained from compounds is shown below:

 

The workflow for using SimFinder is as follows:

Validation Results

• ROC curve

• We compare the top results obtained using our similarity search method with a fingerprint-based similarity search.

More details on this technology can be found here.

References

1. Huahai He; Ambuj K. Singh; Closure-tree: An Index Structure for Graph Queries. Proceedings of the 22nd International Conference on Data Engineering (ICDE), April, 2006, pp 38 - 50 DOI :10.1109/ICDE.2006.37.