(567b) Computer Aided Generation of Diverse and Synthesizable Molecular Library

Modern drug discovery is extending into chemical spaces with increased complexity. It is desirable to explore these spaces more efficiently. In order to maximize the information gained from testing new molecules, the concept of molecular library has been widely adopted in medicinal chemistry.

Molecular libraries are collections of compounds that have, or expected to have, similar structures or properties. Accessing all molecules in a molecule library can enable the rapid generation of structure-activity relationships during the hit-to-lead or lead optimization phases, which can greatly enhance the efficiency of drug discovery. Small molecule libraries can be constructed in many different ways. A library may be a collection of compounds extracted from the literature with similar functionality;^1,2 or it may be designed by enumerating possible side groups as decorations of a common core scaffold;^3–5 it may also be designed using any one of the increasing number of in silico tools for the generation of drug-like compound libraries.^6–8

Despite the significant progress in molecular library generation, there is a usually a tradeoff between accessibility and diversity – libraries that are constructed by simply substituting functional groups on the same scaffold are usually limited in diversity, and libraries with larger structural diversity tend not to be all accessible through similar reactions. In this work, we leverage informatics tools to maximize the diversity of molecular libraries as well as their synthetic accessibility. Starting from a given molecule, using retrosynthesis analysis and reaction prediction models, we can generate libraries that can be accessed through alternative late-stage functionalization. The diversity of the molecular libraries is assessed by average pairwise similarity, so that the most diverse molecular library can be identified. This work provides insight on expanding the exploration of surrounding chemical space of a molecule with minimal additional effort.

1 U. F. Röhrig, S. R. Majjigapu, P. Vogel, V. Zoete and O. Michielin, J. Med. Chem., 2015, 58, 9421–9437.

2 B. T. Xin, E. M. Huber, G. De Bruin, W. Heinemeyer, E. Maurits, C. Espinal, Y. Du, M. Janssens, E. S. Weyburne, A. F. Kisselev, B. I. Florea, C. Driessen, G. A. Van Der Marel, M. Groll and H. S. Overkleeft, J. Med. Chem., 2019, 62, 1626–1642.

3 K. Kannan Sivaraman, A. Paiardini, M. Sieńczyk, C. Ruggeri, C. A. Oellig, J. P. Dalton, P. J. Scammells, M. Drag and S. McGowan, J. Med. Chem., 2013, 56, 5213–5217.

4 R. Fleeman, T. M. Lavoi, R. G. Santos, A. Morales, A. Nefzi, G. S. Welmaker, J. L. Medina-Franco, M. A. Giulianotti, R. A. Houghten and L. N. Shaw, J. Med. Chem., 2015, 58, 3340–3355.

5 G. M. Keseru, D. A. Erlanson, G. G. Ferenczy, M. M. Hann, C. W. Murray and S. D. Pickett, J. Med. Chem., 2016, 59, 8189–8206.

6 R. Gómez-Bombarelli, J. N. Wei, D. Duvenaud, J. M. Hernández-Lobato, B. Sánchez-Lengeling, D. Sheberla, J. Aguilera-Iparraguirre, T. D. Hirzel, R. P. Adams and A. Aspuru-Guzik, ACS Cent. Sci., 2018, 4, 268–276.

7 M. H. S. Segler, T. Kogej, C. Tyrchan and M. P. Waller, ACS Cent. Sci., 2018, 4, 120–131.

8 A. Gupta, A. T. Müller, B. J. H. Huisman, J. A. Fuchs, P. Schneider and G. Schneider, Mol. Inform., 2018, 37, 1700111.