Abstract
BackgroundAmong the various molecular fingerprints available to describe small organic molecules, extended connectivity fingerprint, up to four bonds (ECFP4) performs best in benchmarking drug analog recovery studies as it encodes substructures with a high level of detail. Unfortunately, ECFP4 requires high dimensional representations (≥ 1024D) to perform well, resulting in ECFP4 nearest neighbor searches in very large databases such as GDB, PubChem or ZINC to perform very slowly due to the curse of dimensionality.ResultsHerein we report a new fingerprint, called MinHash fingerprint, up to six bonds (MHFP6), which encodes detailed substructures using the extended connectivity principle of ECFP in a fundamentally different manner, increasing the performance of exact nearest neighbor searches in benchmarking studies and enabling the application of locality sensitive hashing (LSH) approximate nearest neighbor search algorithms. To describe a molecule, MHFP6 extracts the SMILES of all circular substructures around each atom up to a diameter of six bonds and applies the MinHash method to the resulting set. MHFP6 outperforms ECFP4 in benchmarking analog recovery studies. By leveraging locality sensitive hashing, LSH approximate nearest neighbor search methods perform as well on unfolded MHFP6 as comparable methods do on folded ECFP4 fingerprints in terms of speed and relative recovery rate, while operating in very sparse and high-dimensional binary chemical space.ConclusionMHFP6 is a new molecular fingerprint, encoding circular substructures, which outperforms ECFP4 for analog searches while allowing the direct application of locality sensitive hashing algorithms. It should be well suited for the analysis of large databases. The source code for MHFP6 is available on GitHub (https://github.com/reymond-group/mhfp).
Highlights
Many uses of cheminformatics require the quantification of the similarity between molecules
Fingerprint design The MinHash fingerprint (MHFP) described combines the concept of extended connectivity used for extended connectivity fingerprint (ECFP) with MinHash as a hashing scheme to later enable locality sensitive hashing (LSH)-based approximate nearest neighbor search (ANN) searches
2048-D MHFP6 was comparable to 16,384-D ECFP4, it still performed better in terms of BEDROC20 and RIE20. 2048-D MHFP6 performed significantly better in area under the curve (AUC) than 2048-D MHECFP4 while non-significantly better in EF1, EF5, BEDROC100 and RIE100 and worse in BEDROC20 and BEDROC100
Summary
Many uses of cheminformatics require the quantification of the similarity between molecules. Among the assortment of fingerprints for the comparison of molecules in use today, extended connectivity fingerprint (ECFP) is the most prominent due to its outstanding performance in molecular structure comparisons requiring the identification of compounds with similar bioactivity, as assessed in benchmarking studies [6, 7]. The performance of ECFP results from a precise encoding of molecular structure, which is achieved by using high-dimensional vectors, typically d ≥ 1024 , with the consequence that linear searching becomes slow when applied to very large databases such as GDB, PubChem or ZINC [8,9,10]. ECFP4 requires high dimensional representa‐ tions (≥ 1024D) to perform well, resulting in ECFP4 nearest neighbor searches in very large databases such as GDB, PubChem or ZINC to perform very slowly due to the curse of dimensionality
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
