Spot-on Biomolecular NMR with Deep Learning

We will revise methodology of biomolecular NMR to meet new challenges and opportunities of the post AlphaFold2 integrative life-science studies.

We develop a task-oriented approach that combines a novel set of selective and spin system optimized NMR experiments with a next-generation signal processing and analysis using deep learning, artificial neural networks, and automation.The new methodology, which we call Focused Spectroscopy (FOSY), deviates from the traditional approach of uniform data sampling and universal broadband experiments.

Instead, FOSY “focuses” sensitivity and resolution at a few but the most important spins residing at hotspots of a protein system.

In its turn, advent of the deep learning and artificial neural networks will significantly rehaul the NMR toolbox for spectra acquiring, processing, and analysis.

The new N-FOSY methodology will be developed, finetuned, and demonstrated on a number of collaboration projects on biomedically important and challenging protein systems including the T-cell receptors and peptide bound major histocompatibility complexes (TCR-pMHC), 44 kDa mucosa-associated lymphoid tissue lymphoma translocation protein 1 (MALT1), 57 kDa photosensory module of bacterio-phytochrome, and the 441aa intrinsically disordered protein Tau.

The innovative N-FOSY methodology and the novel results afforded by its application in our showcase studies will greatly extend the usability of NMR and enable a broad range of new challenging research lines.