Dataset acquisition and curation are often the hardest and most time-consuming parts of a machine learning endeavor. This is especially true for proteomics-based LC-IM-MS datasets, due to the high-throughput data structure with high levels of noise and complexity between raw and machine learning-ready formats. While predictive proteomics is a field on the rise, when predicting peptide behavior in LC-IM-MS setups, each lab often uses unique and complex data processing pipelines in order to maximize performance, at the cost of accessibility and reproducibility. For this reason we introduce ProteomicsML, an online resource for proteomics-based datasets and tutorials across most of the currently explored physicochemical peptide properties. This community-driven resource makes it simple to access data in easy-to-process formats, and contains easy-to-follow tutorials that allow new users to interact with even the most advanced algorithms in the field. ProteomicsML provides datasets that are useful for comparing state-of-the-art (SOTA) machine learning algorithms, as well as providing introductory material for teachers and newcomers to the field alike. The platform is freely available on https://www.proteomicsml.org/ and we welcome the entire proteomics community to contribute to the project at https://github.com/proteomicsml/.
Supplementary Table 1
Proteomics ML publications along with links to the ProteomeXchange datasets used for training or testing.
Supplementary Table 2
Public ProteomeXchange datasets that have been used for ML training or benchmarking.