Employing Ensemble Protein-Ligand Interaction Fingerprints to Mimic Induced-Fit Theory in Structure-Based Virtual Screening Targeting Dipeptidyl Peptidase IV

Enade Perdana Istyastono (Department of Pharmacy, Faculty of Medicine and Health Sciences, Atma Jaya Catholic University of Indonesia, Jalan Pluit Raya 2, Jakarta 14440, Indonesia)
Bonifacius Ivan Wiranata (Pharmaceutical Sciences Department, Faculty of Pharmacy, Widya Mandala Catholic University, Surabaya, Indonesia)
Florentinus D.O. Riswanto (Faculty of Pharmacy, Sanata Dharma University, Yogyakarta 55282, Indonesia)
Fransiska Kurniawan (School of Pharmacy, Bandung Institute of Technology, Jalan Ganesha 10, Bandung 40132, Indonesia)
Tasia Amelia (School of Pharmacy, Bandung Institute of Technology, Jalan Ganesha 10, Bandung 40132, Indonesia)
Nunung Yuniarti (Department of Pharmacology and Clinical Pharmacy, Faculty of Pharmacy, Universitas Gadjah Mada, Yogyakarta 55281, Indonesia)
Eko Adi Prasetyanto (Department of Pharmacy, Faculty of Medicine and Health Sciences, Atma Jaya Catholic University of Indonesia, Jalan Pluit Raya 2, Jakarta 14440, Indonesia)

Journal of Pharmaceutical Sciences and Community
Vol. 23 No. 1 (2026)

We have successfully employed PyPLIF HIPPOS in retrospective Structure-Based Virtual Screening (SBVS) campaigns targeting some G-protein coupled receptors (GPCRs), which could pinpoint the molecular determinants of the protein-ligand bindings and increase the quality of the SBVS protocols. We were then tempted to append with molecular dynamics simulations using YASARA-Structure to mimic the induced-fit theory in the construction of SBVS protocols targeting dipeptidyl peptidase IV (DPP4). The protocol was retrospectively validated by employing the DPP4 ligands and decoys provided by the Directory of Useful Decoys: Enhanced (DUDE). The best SBVS protocol from this research has the balanced accuracy (BA) value of 0.836, which could be used further in prospective screening campaigns.