An End-to-end Deep Learning Method For Mass Spectrometry Data Analysis To Reveal Disease-specific Metabolic Profiles

DeepMSProfiler is an ensemble deep learning method that accurately differentiates between healthy individuals and patients with lung nodules or lung adenocarcinoma using metabolomic data. It also effectively removes batch effects and identifies unannotated metabolites associated with specific classifications. The method has potential applications in disease diagnosis and mechanism discovery.