Inverse molecular docking fingerprints for target identification

AUTOR(I) / AUTHOR(S): Marko Jukić, Vid Ravnik, Urban Bren

Download Full Pdf   

DOI:  10.46793/ICCBIKG25.075J

SAŽETAK / ABSTRACT:

Inverse molecular docking fingerprinting (IMDF) extends classical inverse docking by combining large-scale docking against thousands of experimentally resolved human protein binding sites with a standardized, quantitative representation of ligand–target interaction patterns. Namely, each compound is docked into a comprehensive, highly non-redundant database of protein binding sites and its docking scores are transformed into a multidimensional fingerprint reflecting the binding potential. Similarity between fingerprints can then be quantified and subjected to clustering experiments. This enables the identification of high-scoring potential targets enriched across multiple ligands and distinct groups of compounds sharing comparable binding profiles. Moreover we postulate structural or physicochemical features most likely responsible for target engagement can be effectively inferred. As application examples; drug repurposing against SARS-CoV-2 and binding of phytocannabinoids, IMDF successfully recovered known targets, revealed novel potential therapeutic proteins and clustered chemically diverse ligands into interpretable structural classes reflecting common modes of action. By decoupling ligand clustering from direct chemical similarity, IMDF provides a powerful cheminformatics tool to map chemical space and guide the prioritization of underexplored molecules or scaffolds for follow-up studies. This makes it especially valuable for elucidating mechanisms of action and discovering new protein–ligand interaction profiles.

KLJUČNE REČI / KEYWORDS:

cheminformatics, inverse molecular docking, inverse molecular docking fingerprints (IFMD), small-molecule protein interactions, in-silico drug design, drug repurposing, mode-of-action (MOA)

PROJEKAT / ACKNOWLEDGEMENT:

Financial resources through the Slovenian Research and Innovation Agency (ARIS) project and program grants P2-0438, I0-E015, P1-0403, Z1-50021, L2-4430, J1- 4414, J3-4497, J7-4638, J3-4498, J1-4398, J4-4633, J7-50043, J1-50034, GC-0001, J1-60001, J2-60044, and L7-60161.

LITERATURA / REFERENCES:

• Chen YZ, Ung CY. Prediction of potential toxicity and side effect protein targets of a small molecule by a ligand–protein inverse docking approach. Chem Res Toxicol. 14 (2001) 398–405.
• Rollinger JM, Schuster D, Danzl B, et In silico target fishing for natural products by inverse docking. Nat Prod Rep. 26 (2009) 965–983.
• Furlan V, Kores K, Bren U. Mechanistic insights into adverse side effects of natural compounds using inverse docking. Front Pharmacol. 9 (2018) 1193.
• Kores K, Furlan V, Lešnik S, et al. Inverse docking for identifying novel targets of natural products. Front Pharmacol. 12 (2021) 678543.
• Kores K, Furlan V, Bren U. Target fishing of plant-derived compounds by inverse docking. Front Pharmacol. 13 (2022) 874562.
• Fine J, Konc J, Janežič D, et ProBiS-Dock: structural comparison and docking in large protein binding-site databases. J Chem Inf Model. 60 (2020) 5294–5302.
• Konc J, Lešnik S, Kores K, et al. The ProBiS-Dock database of non-redundant protein–ligand binding sites. Nucleic Acids Res. 50 (2022) D1497–D1503.
• Sheridan RP, McMasters DR, Tibbitts J. Early recognition metrics in virtual screening. J Chem Inf Comput Sci. 41 (2001) 1395–1400.
• Truchon JF, Bayly CI. Evaluating virtual screening methods: the good, the bad, and the BEDROC. J Chem Inf Model. 47 (2007) 488–508.
• Empereur-Mot C, Guillemain H, Latouche A, et al. Total gain metric for early recognition. J Chem Inf Model. 55 (2015) 1605–1612.
• Jukič M, Bren U, Konc J, et al. Repurposing of drugs for SARS-CoV-2 using inverse docking fingerprints. Front Chem. 9 (2021) 757826.
• Nielsen F, Nielsen T. Agglomerative hierarchical clustering methods: performance and applications. Pattern Recogn. 53 (2016) 108–118.
• Ravnik V, Jukič M, Furlan V, et al. Uncovering the molecular targets of phytocannabinoids: mechanistic insights from inverse molecular docking fingerprint approaches. Front Pharmacol. 16 (2025)
• Berman HM, Westbrook J, Feng Z, et al. The Protein Data Bank. Nucleic Acids Res. 28 (2000) 235–242.
• Konc J, Janežič D.  ProBiS algorithm for detection of structurally similar protein binding sites. J Chem Inf Model. 50 (2010) 1487–1493.
• Konc J, Janežič D. An improved ProBiS algorithm for large-scale detection of structurally conserved protein binding sites. Bioinformatics. 33 (2017) 134–142.
• Depolli M, Konc J, Janežič D, et Maximum clique algorithm for reconstructing ligands from fragments docked into protein binding sites. J Chem Inf Model. 53 (2013) 3097–3112.