Peak lists, assignments, and structures of 100 proteins determined by the ARTINA algorithm

The time-consuming and complex data analysis process constitutes a limitation for biomolecular NMR. It typically requires weeks or months of manual work of a trained expert to identify and interpret thousands of signals, recorded in a series of multidimensional NMR spectra, in order to determine the sequence-specific resonance assignments or the three-dimensional structure of a single protein. 
To solve this problem, we developed ARTINA, a machine learning-based method that uses as input only NMR spectra and the protein sequence, and delivers signal positions, resonance assignments, and structures strictly without any human intervention. Tested on a 100-protein benchmark comprising 1329 multidimensional NMR spectra, ARTINA demonstrated its ability to solve structures with 1.44 Å median RMSD to the PDB reference and to identify 91.36% correct NMR resonance assignments. 
ARTINA is freely accessible for non-commercial users at NMRtist (https://nmrtist.org), an online platform that combines deep learning, large-scale optimization, and cloud computing to offer full automation of NMR spectra analysis. Our website provides virtual storage for NMR spectra deposition together with a set of applications designed for automated peak picking, chemical shift assignment, and protein structure determination. The system can be used by non-experts and without IT infrastructure on the user's side, allowing protein NMR spectra interpretation within hours after completion of the measurements. With NMRtist, the effort for a protein assignment or structure determination by NMR essentially reduced to the preparation of the sample and the spectrum measurements.