SimFinder

SimFinder is a graph-based similarity searching tool for performing topological searches over graph databases. SimFinder's innovative C-Tree graph indexing technology allows it to answer graph similarity queries extremely efficiently even over millions of graphs. At the same time, it can be easily configured for many search scenarios via user-definable similarity measures, graph attributes, and subgraph abstractions.

SimFinder/Chem       New Leads with Structural Intuition

SimFinder/Chem is a specialized version of SimFinder tailored to cheminformatics. When used to search over large chemical compound databases, SimFinder/Chem can deliver complementary search results when compared with traditional similarity search mechanisms such as fingerprints. This is a direct consequence of SimFinder's novel topological fuzzy graph similarity search method which maintains the full connectivity information of topological fragments present in a compound while allowing for slight variations in the local structure. In addition, SimFinder/Chem points the chemist or pharmacologist to the matched fragments for an easier interpretation of the search results. This can make SimFinder/Chem a valuable complementary tool in the lead discovery toolbox. → SimFinder Quality

 

For illustration, below is an example demonstrating the difference between conventional molecular similarity searching and SimFinder's fuzzy matching technology:

The bottom result shows how SimFinder also returns molecules that have a slight deviation from the query molecule (here additional S and N atoms). The shown results would both be ranked very high. Note also that this type of fuzzy matching can be used both with substructure searches and whole molecule similarity searches.

As another example, consider searching fuzzily for multiple substructures as in the following example:

SimFinder will search for both structures using fuzzy matching and rank the results by how well both matched with each molecule. The result on the right shows some of the top ranked molecules. SimFinder highlights information in the output file that can be rendered in different colors (here, red and blue) in order to show the user which substructure has been matched with which fragment of the molecule.

The output of any query can then be used for compound design, activity knowledge, or as input to visualization tools. SimFinder/Chem supports MOL2 and SDF data formats.