Chemical Library Space: Definition and DNA-Encoded Library Comparison Study Case

The development of DNA-encoded library (DEL) technologyintroducednew challenges for the analysis of chemical libraries. It is oftenuseful to consider a chemical library as a stand-alone chemoinformaticobject represented both as a collection of independent molecules,and yet an individual entity in particular, when they are inseparablemixtures, like DELs. Herein, we introduce the concept of chemicallibrary space (CLS), in which resident items are individual chemicallibraries. We define and compare four vectorial library representationsobtained using generative topographic mapping. These allow for aneffective comparison of libraries, with the ability to tune and chemicallyinterpret the similarity relationships. In particular, property-tunedCLS encodings enable to simultaneously compare libraries with respectto both property and chemotype distributions. We apply the variousCLS encodings for the selection problem of DELs that optimally "match"a reference collection (here ChEMBL28), showing how the choice ofthe CLS descriptors may help to fine-tune the "matching"(overlap) criteria. Hence, the proposed CLS may represent a new efficientway for polyvalent analysis of thousands of chemical libraries. Selectionof an easily accessible compound collection for drug discovery, asa substitute for a difficult to produce reference library, can betuned for either primary or target-focused screening, also consideringproperty distributions of compounds. Alternatively, selection of librariescovering novel regions of the chemical space with respect to a referencecompound subspace may serve for library portfolio enrichment.