Towards Ultimate NMR Resolution with Deep Learning

The dataset contains processed solution-state protein NMR spectra of MALT1 (45 kDa), Azurin (14 kDa), and Tau (disordered, 45.8 kDa), derived from experimentally recorded 2D and 3D data obtained in earlier studies and published previously: (1) DOI: 10.1371/journal.pone.0146496; DOI: 10.1007/s12104-022-10105-3; (2) DOI: 10.1110/ps.0225403; (3) DOI: 10.1002/anie.202102758 All processed data are stored in NMRPipe format (.ft2 and .ft3 files) and were generated using standard NMR processing procedures. The data can be read and visualized using NMRPipe-compatible software, such as NMRPipe (https://www.ibbr.umd.edu/nmrpipe/), the nmrglue Python package (https://github.com/jjhelmus/nmrglue), or other software supporting the NMRPipe format, including CCPN 3.0 (https://ccpn.ac.uk/) and later versions. These processed spectra are used as input files for AI-based methods to improve NMR spectral resolution.