Probing More Of That Vast Chemical Space

Even with their rosters of chemicals sometimes numbering 1 million, screening libraries represent such a minuscule fraction of all possible druglike molecules that it's amazing drug candidates are ever found. By adding chemical scaffolds that now are missing from common screening libraries, more of that vast chemical space can be included, according to a study by Jérôme Hert, Brian K. Shoichet, and coworkers at the University of California, San Francisco (Nat. Chem. Biol., DOI: 10.1038/nchembio.180). They compared commercial screening libraries with chemical space—as represented in the Generated Database (more than 26 million C-, N-, O-, and F-containing compounds of up to 11 atoms). They also compared the libraries with protein-recognizing compounds from two major databases of metabolite and natural products. The study confirms, as might have been expected, that commercial libraries are strongly biased to include many more compounds with biogenic structures than a random sampling of chemical space would have produced. However, hundreds of biogenic ring scaffolds are missing from commercial libraries. These "could be used to further increase the bias in screening libraries toward those molecules that proteins have evolved to recognize," Shoichet and coworkers point out.