Deep Learning-Assisted Peak Curation for Large-Scale LC-MS Metabolomics

Available automated methods for peak detection in untargeted metabolomics suffer from poor precision. We present NeatMS, which uses machine learning based on a convoluted neural network to reduce the number and fraction of false peaks. NeatMS comes with a pre-trained model representing expert knowledge in the differentiation of true chemical signal from noise. Furthermore, it provides all necessary functions to easily train newmodels or improve existing ones by transfer learning. Thus, the tool improves peak curation and contributes to the robust andscalable analysis of large-scale experiments. We show how tointegrate it into different liquid chromatography-mass spectrom-etry (LC-MS) analysis workflows, quantify its performance, and compare it to various other approaches. NeatMS software is available as open source on github under permissive MIT license and is also provided as easy-to-install PyPi and Bioconda packages.

Automated summary

NeatMS is a machine learning tool based on convolutional neural networks that reduces false peaks in untargeted metabolomics. It includes a pre-trained model and functions for training new models, improving peak curation, and facilitating the analysis of large-scale experiments.