Supplementary materials for NMRcraft publication: NMR, Chromatograms, 3D models, Videos and Script

This repository contains supplementary materials associated with the manuscript: Automated NMRcraft Deconvolution Enables Hyperspace-Wide Quantification of Crude Reaction Mixtures The dataset includes experimental data collected during the investigation of Chan–Lam coupling reactions and their sensitivity to several reaction parameters, including the stoichiometry of boronic acid, amine, copper source and base, as well as temperature, reaction duration and the use of an oxygen atmosphere instead of air. Four substrate families were investigated: morpholine, 7-azaindole, 5-bromo-7-azaindole, and 1H-indazole. The resulting NMR measurements were used to evaluate the performance of NMRcraft, a Python-based algorithm developed for the analysis of crude reaction mixtures by NMR spectroscopy. NMRcraft uses libraries of reference spectra to identify and quantify mixture components while accounting for peak shifts and broadening. The workflow is compatible with low-field NMR data and spectra acquired in non-deuterated solvents. The repository contains: Two videos: one demonstrating automated benchtop NMR measurements, and one illustrating oxygen sparging across a full 54-well plate. 3D-printing/CAD files to reproduce the head connector used to sparge 54 vials simultaneously with oxygen. 50 chromatograms of crude reaction mixtures, together with the calibration curve used to quantify N-phenyl-7-azaindole, provided in PDF format. An MNova script used to process repositories containing raw NMR FID files and generate CSV files of NMR spectra after solvent-based chemical shift correction, phase correction, and baseline correction. NMRcraft-environnement.yml, which can be used to create a Python environment compatible with NMRcraft. NMRcraft.py, a Python script used to analyse the CSV spectra provided in the Data folder and determine the composition of crude reaction mixtures using the supplied reference spectral libraries. Four substrate-specific data folders, named Data-X, where X corresponds to the substrate family. These folders contain the NMR data used for NMRcraft analysis and are organized by substrate: morpholine, 7-azaindole, 5-bromo-7-azaindole, and 1H-indazole. Each Data-X folder contains two main subfolders: ReferencesThis folder contains the raw FID files and CSV spectra of the pure reference analytes used by NMRcraft to build the reference spectral library. It also contains a Calibration folder, which includes a series of measurements of one analyte at different concentrations. These calibration spectra are used to generate the calibration curve required for quantification. DatasetThis folder contains the experimental reaction datasets corresponding to 1575 reactions. It is divided into two main subfolders: Hyperspace CSV files, which contains CSV files compiling the yields or compositions obtained for each analyte as a function of the reaction conditions tested. Robot_Run, which contains the raw FID files and CSV spectra of the crude reaction mixtures. When applicable, this folder also contains a dedicated subfolder for measurements performed with an internal standard, used to determine ground-truth quantification values. The hyperspace CSV files can be directly visualised using HyperspaceViewer.exe, provided in: Jia, Y., Frydrych, R., Sobolev, Y. I. et al. Robot-assisted mapping of chemical reaction hyperspaces and networks. Nature 645, 922–931 (2025). https://doi.org/10.1038/s41586-025-09490-1