deepPseudoMSI

deepPseudoMSI converts raw liquid chromatography–mass spectrometry (LC-MS) data into pseudo-mass spectrometry images and applies deep learning to improve untargeted metabolomics analyses for disease diagnosis and precision medicine.

Key Features:

• Raw LC-MS to pseudo-MS image conversion: Converts raw LC-MS-based untargeted metabolomics data into pseudo-mass spectrometry images for downstream analysis.
• Deep learning-based analysis: Processes pseudo-mass spectrometry images using deep learning algorithms to extract diagnostic features.
• Addressing metabolite identification challenges: Targets issues inherent to LC-MS untargeted metabolomics, including metabolite identification limitations.
• Information retention and enhancement: Retains and enhances critical information that may be lost in standard metabolomics workflows.
• Improved reproducibility and robustness: Enhances reproducibility and reliability of metabolomic analyses compared with traditional methods.
• Individualized diagnostic capability: Supports precise individualized diagnoses by leveraging image-based deep learning on metabolomic data.

Scientific Applications:

• Precision medicine: Supports metabolic-based precision medicine by enabling individualized diagnostic analyses from LC-MS data.
• Disease diagnosis: Improves disease diagnosis using untargeted LC-MS metabolomic profiles converted to pseudo-mass spectrometry images.
• Systematic metabolic profiling: Facilitates systematic metabolic profiling in research and clinical contexts using LC-MS-based untargeted metabolomics.

Methodology:

Converts raw LC-MS data into pseudo-mass spectrometry images, processes those images using deep learning algorithms, and evaluates performance on real datasets.