Mass Spectrometry Techniques and Applications

The paper characterizes an Omnitrap-Orbitrap mass spectrometer with multiple ion activation methods, showing its versatility in peptide and protein analysis.

SmartGate improves spatial segmentation and biomarker identification in IMS data through iterative peak picking and graph attention auto-encoder modeling.

The paper uses a machine learning-based high-throughput approach to analyze charge state distributions of peptides, revealing two distinct charging regimes with different sequence determinants.

The paper evaluates pixel-cell deconvolution methods in spatial single-cell metabolomics using fluorescein as ground truth, showing that weighted average outperforms linear inverse modeling and demonstrating the impact of ion suppression.

KineticMSI quantifies spatial isotopic incorporation in kMSI data across replicated experiments, revealing metabolic changes in HD mouse hippocampus.

TIMSCONVERT converts trapped ion mobility data to open formats for compatibility with downstream analysis tools.

This paper develops an empirical Bayes functional model for HDX-MS data to improve statistical power and reduce false positives by incorporating a physical model of hydrogen deuterium exchange.

MSNovelist generates small molecule structures de novo from mass spectra, outperforming database-based methods for novel compounds.

The paper creates a large-scale CCS database for drugs and metabolites and develops a machine learning model to predict CCS values for different molecular forms.

The paper introduces a post-acquisition method to correct mass drift in DESI-MSI data, reducing mass errors by up to 60 ppm and improving molecular annotation and spatial distribution quality.

The paper compares LSTM, CNN, and transformer models for predicting retention time and collisional cross section in mass spectrometry data, finding LSTM and CNN perform similarly while the transformer underperforms.

massNet uses deep learning to classify spatial mass spectrometry data without preprocessing, achieving higher accuracy and faster speed than classical methods for tumor delineation in mouse brain tissue.

This paper demonstrates that Trapped Ion Mobility Spectrometry (TIMS) reduces cofragmentation rates in mass spectrometry-based proteomics, increasing peptide identification rates.

MassGenie uses a transformer-based deep learning model to predict small molecule structures from mass spectra, achieving accurate results without explicit fragmentation rules.

This paper introduces moleculaR, a computational framework for probabilistic mapping and statistical testing of metabolites in mass spectrometry imaging data.

This paper introduces a modified mass spectrometry system for depositing gas-phase protein complexes onto EM grids, enabling 3D structure determination of non-covalent protein complexes like GroEL.

The paper introduces msiPL, a deep learning model for analyzing MSI data, enabling efficient non-linear visualization and identification of biologically relevant patterns.

The paper introduces PyHDX, a tool for analyzing HDX-MS data to compute residue-level Gibbs free energy across protein conditions and homologues.

molDiscovery improves small molecule identification by learning fragmentation patterns from mass spectra, achieving higher efficiency and accuracy than existing methods.

The paper develops a sequence-specific model to predict peptide collision cross-section values using amino-acid sequence and charge state information from proteomic data.

The paper introduces Madi, a workflow combining MALDI-MSI and machine learning to analyze FFPE tissues, revealing metabolic patterns and pathological features in human tissues.

This paper introduces an automated workflow using SHAP to identify biomarker candidates in imaging mass spectrometry data by ranking molecular species based on their predictive importance for biological class recognition.

MS2DeepScore uses a Siamese neural network to predict structural similarity from mass fragmentation spectra, outperforming classical methods and enabling chemically meaningful spectral embeddings.

The paper couples SPR imaging with MALDI MS to identify RNA-binding proteins from yeast mitochondrial lysate, revealing new cofactors for intron splicing beyond Mss116.

pDeep3 improves spectrum prediction accuracy using few-shot learning with minimal computational resources across different mass spectrometry instruments.

This paper describes a method to measure A2M concentration in human serum using MALDI mass spectrometry by analyzing the release of a specific peptide after trypsin interaction.

The paper introduces in vivo subcellular mass spectrometry to analyze proteo-metabolomic profiles in Xenopus laevis embryos, revealing differential quantitative activities between cell types.

This paper uses deep learning to improve clustering of mass spectrometry imaging data by transforming ion images into neural ion images, resulting in more consistent and fine-grained spatial clusters.

The paper measures CCS values for a million peptides across five organisms and develops a method to predict CCS for any peptide and organism.

This paper improves volatile metabolome analysis using SESI-Orbitrap MS by minimizing ion competition through optimized m/z window splitting.