Benchmarking MS/MS featurization strategies for machine learning-driven metabolite structure annotation - Dataset

This repository serves as companion to the upcoming article to be published in JASMS, which focuses on comparing different strategies for MS/MS featurization in order to train ML models used in molecular structure annotation tasks. Please refer to the README.md file for more information about how this submission is structured: datasets, models, how to reproduce the results, etc. In order to replicate the study, all the zip files contained in this repository must be extracted into a folder named Data, which should be placed inside the benchmark_ms2_featurization folder obtained by cloning the Github repository https://github.com/RogerGinBer/benchmark_ms2_featurization. Then, follow the instructions contained in that repository README.md file.