MUMBAI, India, Jan. 23 -- Intellectual Property India has published a patent application (202511126501 A) filed by Swami Rama Himalayan University, Dehradun, Uttarakhand, on Dec. 14, 2025, for 'a system for neural similarity-gating pre-processing in drug-target binding prediction pipelines.'
Inventor(s) include Dr. Deepak Srivastava; Sameer Rajesh Chavan; Dr. Pramod Kumar; and Dr. Suman Pant.
The application for the patent was published on Jan. 23, under issue no. 04/2026.
According to the abstract released by the Intellectual Property India: "The present invention discloses a neural similarity-gating pre-processor configured as an upstream decision layer in computational drug-target interaction pipelines. Candidate drug molecules represented as SMILES strings, molecular graphs or 3D conformers are encoded into latent-space embeddings using a neural molecular encoder. Target-specific features of a protein, including sequence- and structure-derived descriptors, are jointly embedded and used to condition a similarity-gating network that computes neural relevance scores for each molecule. An adaptive, target-aware threshold is then derived from learned statistics over known actives and inactives, and a pre-binding decision controller selectively forwards only molecules exceeding the threshold to downstream docking, binding-affinity prediction or virtual-screening engines. Low-relevance molecules are discarded prior to expensive computation, thereby reducing CPU/GPU load, improving hit enrichment and increasing predictive fidelity. The invention is implementation-agnostic and integrates with existing docking or AI-based drug discovery platforms as a modular pre-processing component."
Disclaimer: Curated by HT Syndication.
