AI platform reconstructs molecular structures from NMR data

Researchers have developed a novel platform called Inverse-IMPRESSION, utilizing an inverted Graph Transformer Network named IMPRESSION-G2, to reconstruct molecular structures from experimental Nuclear Magnetic Resonance (NMR) spectroscopic data. This system employs a multi-stage process involving bond connectivity prediction, structure correction, and noise-augmented multi-shot prediction to identify the most fitting molecular structure. The platform demonstrated success in identifying structures for 77.8% of molecules with up to 30 heavy atoms using simulated NMR data and 53% of molecules using experimental data, marking a significant advancement in automated molecular structure elucidation. AI

IMPACT This research advances the application of graph-based machine learning in chemistry, potentially accelerating drug discovery and materials science.

RANK_REASON The cluster describes a scientific paper detailing a new method for molecular structure elucidation using graph-based machine learning. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Graph Transformer Networks
• IMPRESSION-G2
• Inverse-IMPRESSION

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →