Laccase-catalyzed oxidation of aromatic substrates

These data, published in line with reproducibility standards, accompany the paper "Prediction of laccase-catalyzed oxidation of aromaticsubstrates using machine learning" by Yulia Kulagina, Christian Goldhahn, Ramon Weisshaupt and Mark Schubert. Results of the oxidation of various aromatic substrate chemical groups were obtained through experimental investigations conducted at the Swiss Federal Laboratories for Materials Science and Technology (Empa). The experiments were carried out according to the protocol described by Reiss et al. 2013 [Reiss, R., Ihssen, J., Richter, M., Eichhorn, E., Schilling, B. and Thöny-Meyer, L., 2013. Laccase versus laccase-like multi-copper oxidase: a comparative study of similar enzymes with diverse substrate spectra. PloS one, 8(6), p.e65633.], using three different laccase enzymes: Trametes versicolor (f-tve), Myceliophthora thermophila (f-mth), and Bacillus pumilus (bpu-lac). The experimental datasets contain 250, 190 and 172 records for the f-tve, f-mth and bpu-lac laccases, respectively. These datasets provide valuable information on the oxidation profiles and efficiencies of different laccase enzymes when exposed to a variety of aromatic substrates. Each of the two folders contains csv files for the 3 laccases. The dragon-descroptors folder contains the data, containing a large number of molecular descriptors calculated using the DRAGON software. These data were used for training and testing the classical classifiers. These data were used for training and testing the classical classifiers. The smiles folder contains IUPAC names of the substrates and the corresponding SMILES strings (sourced from the WolframAlpha engine), which were used for fine-tuning and testing the ChemBERTa model. Each data point includes:• The chemical structure of the substrate (in either a descriptor-based or a SMILES form),• A binary target label indicating whether oxidation was observed under the specific laccase’s catalytic action.These datasets provide a unique resource for modeling laccase-mediated oxidation, given their experimental origin and specificity to well-characterized enzymes – features often lacking in public databases.