SMIfp (SMILES fingerprint) Chemical Space for Virtual Screening and Visualization of Large Databases of Organic Molecules

J. Schwartz,M. Awale,J. Reymond
DOI: https://doi.org/10.1021/ci400206h
IF: 6.162
2013-07-30
Journal of Chemical Information and Modeling
Abstract:SMIfp (SMILES fingerprint) is defined here as a scalar fingerprint describing organic molecules by counting the occurrences of 34 different symbols in their SMILES strings, which creates a 34-dimensional chemical space. Ligand-based virtual screening using the city-block distance CBD(SMIfp) as similarity measure provides good AUC values and enrichment factors for recovering series of actives from the directory of useful decoys (DUD-E) and from ZINC. DrugBank, ChEMBL, ZINC, PubChem, GDB-11, GDB-13, and GDB-17 can be searched by CBD(SMIfp) using an online SMIfp-browser at www.gdb.unibe.ch. Visualization of the SMIfp chemical space was performed by principal component analysis and color-coded maps of the (PC1, PC2)-planes, with interactive access to the molecules enabled by the Java application SMIfp-MAPPLET available from www.gdb.unibe.ch. These maps spread molecules according to their fraction of aromatic atoms, size and polarity. SMIfp provides a new and relevant entry to explore the small molecule chemical space.
What problem does this paper attempt to address?