Aiffinity offers an AI‑enhanced NMR platform that accelerates fragment‑based hit discovery and enables high‑resolution structure determination of proteins up to 30 kDa, including intrinsically disordered targets. Their deepHitExplorer tool combines ultra‑fast photo‑CIDNP NMR with machine‑learning models to predict binding epitopes and select diverse, soluble compounds, while the 4D‑GraFID pipeline automates chemical shift assignment and 4D NMR structure calculation using standard isotopic labeling. The solution is delivered via a SaaS web portal, containerized software, or consulting services for pharmaceutical, biotech, and academic research teams.
Funding
Funding not disclosed
Founders
Product
Problem
Fragment-based drug discovery and protein structure determination rely on labor-intensive NMR experiments that are limited to relatively small, isotopically labeled proteins, making it difficult to target larger or intrinsically disordered proteins and to rapidly identify and optimize hits.
Solution
Aiffinity combines advanced NMR techniques with artificial intelligence to streamline both hit discovery and protein structure elucidation. Their deepHitExplorer platform integrates ultra‑fast photo‑CIDNP ligand‑observed NMR, other biophysical assays, and explainable AI models to predict binding affinity, ligand epitopes, and select diverse, soluble compound sets without requiring a receptor structure. For structural biology, the 4D‑GraFID pipeline uses 4D NMR data and machine‑learning‑driven chemical shift assignment to determine high‑resolution structures of proteins up to 30 kDa, including intrinsically disordered targets, using only standard uniform ¹H,¹³C,¹⁵N labeling. The workflow is offered as a SaaS web portal, containerized software, or consulting service, enabling end‑to‑end gene‑to‑structure pipelines, molecular dynamics refinement, and lead‑optimization tools such as deepScaffOpt.
Target Audience
Primary customers are pharmaceutical and biotech research teams engaged in fragment‑based screening, structure‑based drug design, and the study of intrinsically disordered or larger proteins, as well as academic labs requiring high‑throughput NMR structure determination.
Features
- ultra‑fast photo‑CIDNP 1D NMR screening combined with deep learning for ligand epitope prediction (deepHitExplorer)
- multi‑objective genetic algorithm for diverse, high‑solubility hit selection
- 4D‑GraFID pipeline for automated backbone and full‑atom chemical shift assignment and structure determination of proteins up to 30 kDa using only uniform labeling
- integration of OpenFold and Hamiltonian Replica Exchange MD to refine NMR‑derived structures
- SaaS web portal and containerized versions for local deployment, supporting both hit discovery and structure workflows
- optional orthogonal validation assays (MST, SPR) and advanced molecular simulation consulting (HT‑REX) with Bayesian restraint modeling
- deepScaffOpt deep‑learning ranking algorithm that incorporates ligand‑ and protein‑observed NMR data for lead optimization across diverse chemical series