AI-designed high-resolution 14N solid-state NMR for characterization of biological molecules (AIHANCE-NMR)

The description of the 3D structures of biomolecules is essential for understanding disease mechanisms and developing new drugs. Solid-state NMR is well-suited to characterize large biological systems such as fibrils, viral capsids, and membrane proteins, which are often associated with the development of various diseases. However, calculating high-quality structures requires high-resolution NMR spectra, the acquisition of which traditionally involves expensive and complex isotopic labeling of samples.

The aim of this project is to develop high-sensitivity and high-resolution solid-state NMR methods for studying biological molecules in their natural form, without the need for special isotopic labeling. The project will employ artificial intelligence to design advanced NMR methods that allow clear detection of proton-nitrogen spin-pair signals, which act as a fingerprint of protein structure. If successful, this project will transform the way biomolecules are studied, providing new insights into their structure and properties and paving the way for the characterization of biological systems that have so far remained inaccessible.

Project progress:

January 1st 2026 – March 31st 2026

In the first quarter of the project, the main focus was on optimizing the simulation procedure required for numerical pulse optimization. Several quantum spin dynamics simulation tools were evaluated, including SIMPSON, SLEEPY, and in-house-developed packages, with the aim of identifying the most efficient strategy for generating synthetic spectra.

In the second quarter, using the optimized simulation strategy, we began optimizing pulse waveforms for selective excitation of 14N resonances. During this period, we obtained pulses capable of selectively excitng 14N resonances and generating high-resolution spectra for spins subject to quadrupolar interactions with parameters relevant to biological molecules. Further optimization of pulses capable of yielding high-resolution and high-sensitivity 14N spectra is currently underway.