D3Similarity: A Ligand-Based Approach for Predicting Drug Targets and for Virtual Screening of Active Compounds Against COVID-19

Zhengdan Zhu,Xiaoyu Wang,Yanqing Yang,Xinben Zhang,Kaijie Mu,Yulong Shi,Cheng Peng,Zhijian Xu,weiliang zhu

DOI: https://doi.org/10.26434/chemrxiv.11959323.v1

2020-01-01

Abstract:Discoveringefficient drugs and identifying target proteins are still an unmet but urgent needfor curing COVID-19. Protein structure based docking is a widely appliedapproach for discovering active compounds against drug targets and forpredicting potential targets of active compounds. However, this approach hasits inherent deficiency caused by, e.g., various different conformations withlargely varied binding pockets adopted by proteins, or the lack of true targetproteins in the database. This deficiency may result in false negative results.As a complementary approach to the protein structure based platform forCOVID-19, termed as D3Docking in our recent work, we developed the ligand-basedmethod, named D3Similarity, which is based on the molecular similarityevaluation between the submitted molecule(s) and those in an active compounddatabase. The database is constituted by all the reported bioactive molecules againstthe coronaviruses SARS, MERS and SARS-CoV-2, some of which have target ormechanism information but some don’t. Based on the two-dimensional andthree-dimensional similarity evaluation of molecular structures, virtualscreening and target prediction could be performed according to similarity rankingresults. With two examples, we demonstrated the reliability and efficiency of D3Similarityfor drug discovery and target prediction against COVID-19. D3Similarity isavailable free of charge at https://www.d3pharma.com/D3Targets-2019-nCoV/D3Similarity/index.php.