HerbalDB 2.0: Optimization of construction of three-dimensional chemical compound structures to update Indonesian medicinal plant database

Objective: Development of novel drugs is an important challenge in the pharmaceutical world and industry. In-silico methods are often considered in refinement / correction processes of drug design because they may lower the costs. The in-silico drug discovery process requires a three- Dimensional Structure (3DS) of the chemical compounds as input. Computational 3DSs often exhibit structural mismatches thus affecting the validity of the in-silico drug design process. In a previous study, a 3DS database with 1405 of Indonesian herbal compounds was developed, named HerbalDB. In this database, various structural mismatches were identified in some of the 3DSs. Our study aimed to identify and correct the structural mismatches in the herbalDB and to determine the best method in creating correct 3DS of chemical compounds. Methods: Structural mismatches in the herbal database were identified by molecular visualization. Results: The identification process yielded 170 compounds with structural mismatches that were corrected with 10 different parameters using the MarvinSketch and VegaZZ software, evaluated by molecular visualization. Conclusions: based on 3DS of chemical compound visualization, ∗.mol and ∗.sdf file format created using Dreiding force fields of MarvinSketch are the best method to construct the proper structure of Indonesian medicinal plant's chemical compound database compared with MMFF94, AMBER and CHARMM forcefields.