Objective: Nowadays, many researchers focused on finding and developing the new inhibitor of HDAC4 and HDAC7 using in silico tools such as docking, pharmacophore approaches, and molecular dynamics simulation. The aim of this research is to identify pharmacophore of HDAC4 and HDAC7. Method: In this research, pharmacophore-based virtual screening was used to find new HDAC4 and HDAC7 inhibitor from Indonesian herbal database. From MUBD-HDACs database, active compounds of HDAC4 and HDAC7 were divided into training and test set. Based on pharmacophore model generation for HDAC4 and HDAC7, 10 models were created. All the models were calculated and evaluated using some parameters of validation. Results: The best pharmacophore model for HDAC4 are model 6 and 10, and for HDAC7 is model 1. Pharmacophore model 6 and 10 (HDAC4) have seven pharmacophore features include three HBA, one HBD, one aromatic ring, one negatively ionizable area and 1 hydrophobic. Pharmacophore model 1 (HDAC7) have five pharmacophore features include two HBA, one HBD, one negatively ionizable area and one hydrophobic. These selected models for HDAC4 and HDAC7 were using for virtual screening against Indonesian herbal database. Conclusion: Based on the results of the virtual screening, six hit compounds were obtained such as artocarpesin, avicularin, dimboa glucoside, eriodictin, luteolin and mirabijalone c.
Key words: Pharmacophore, Virtual Screening, HDAC4, HDAC7, Indonesian Herbal Database.