US 12,293,809 B2
Workflow for generating compounds with biological activity against a specific biological target
Aleksandrs Zavoronkovs, Pak Shek Kok (HK); Yan Ivanenkov, Moscow (RU); Daniil Polykovskiy, Moscow (RU); and Aleksandr Aliper, Moscow (RU)
Assigned to INSILICO MEDICINE IP LIMITED, Hong Kong (HK)
Filed by INSILICO MEDICINE IP LIMITED
Filed on Aug. 21, 2020, as Appl. No. 17/000,109.
Claims priority of provisional application 62/891,050, filed on Aug. 23, 2019.
Prior Publication US 2021/0057050 A1, Feb. 25, 2021
Int. Cl. G16C 20/50 (2019.01); C07D 231/56 (2006.01); C07D 245/04 (2006.01); C07D 401/06 (2006.01); C07D 401/12 (2006.01); C07D 403/06 (2006.01); C07D 403/10 (2006.01); C07D 403/12 (2006.01); C07D 409/12 (2006.01); C07D 471/04 (2006.01); C07D 471/10 (2006.01); C07D 487/04 (2006.01); C07D 487/10 (2006.01); C07D 491/20 (2006.01); C07D 513/10 (2006.01); G06N 20/00 (2019.01); G16C 20/40 (2019.01); G16C 20/70 (2019.01); G16H 10/20 (2018.01)
CPC G16C 20/50 (2019.02) [C07D 231/56 (2013.01); C07D 245/04 (2013.01); C07D 401/06 (2013.01); C07D 401/12 (2013.01); C07D 403/06 (2013.01); C07D 403/10 (2013.01); C07D 403/12 (2013.01); C07D 409/12 (2013.01); C07D 471/04 (2013.01); C07D 471/10 (2013.01); C07D 487/04 (2013.01); C07D 487/10 (2013.01); C07D 491/20 (2013.01); C07D 513/10 (2013.01); G06N 20/00 (2019.01); G16C 20/40 (2019.02); G16C 20/70 (2019.02); G16H 10/20 (2018.01)] 16 Claims
 
1. A method, comprising:
receiving input of a biological target or ligand;
receiving input of properties of a generated compound;
receiving at least one generative model trained with reference compounds, the reference compounds identified from one or more groups of compounds based on the biological target or ligand and the properties of the generated compound;
generating structures of generated compounds with each generative model, based on the input of properties of the generated compound, wherein the generated compounds are designed to interact with the biological target and/or correlate with structural features of the ligand, wherein generating includes refining the generated compounds with a rewards function and performing pharmacophore modeling;
prioritizing structures of the generated compounds of each generative model based on at least one reward criteria;
processing prioritized chemical structures of the generated compounds through a Sammon mapping protocol to obtain hit structures, wherein the Sammon mapping protocol uses molecular descriptors applied in the rewards function and a root-mean-square deviation (RMSD) value from the pharmacophore modeling;
removing compounds that cannot be synthesized from the hit structures;
providing chemical structures of the synthesizable hit structures; and
synthesizing the chemical structures of the synthesizable hit structures.