Metabolomics and Mass Spectrometry Studies

This paper introduces DeepSWATH, a deep learning method to improve metabolite identification from SWATH data by reconstructing precursor-ion to MS/MS spectrum associations.

The paper introduces lipidomeR, a tool for analyzing lipidomics data, demonstrating its application in three studies to reveal lipidomic patterns across different conditions.

The paper identifies metabolites in plasma associated with memory decline in middle-aged individuals using metabolomics.

MS-DIAL 4 provides a comprehensive lipidome atlas with retention time, collision cross section, and MS/MS data to standardize and accelerate lipidomics research.

The paper introduces dbnorm, an R package to visualize and correct signal drift in metabolomics data across batches, improving the detection of true biological variation.

The paper describes the fecal metabolome in cryogenically collected samples from healthy participants, identifying metabolites and their variation across different areas of the stool.

MQ is a tool for quantitative analysis of HR-AM data in metabolomic workflows, validated through benchmarking and showcasing utility across different analysis scenarios.

MESSAR recommends metabolite substructures from tandem mass spectra without requiring predefined metabolite candidates.

This paper introduces ncGTW, a method to align LC-MS profiles by detecting and correcting misalignments using compound-specific warping functions and neighbor constraints.

This study evaluates batch correction methods in metabolomics using simulated data, finding that log transformations impact performance and false positives increase with correction.

This paper develops and validates a quantitative NMR-based metabolomics platform for canine serum and plasma, defining reference intervals and demonstrating its clinical relevance.

The paper presents a new workflow using R packages for analyzing metabolite-trait associations, demonstrated on public data and made reproducible.

genuMet identifies and reduces false metabolic features in untargeted metabolomics without pooled QC samples, achieving high true negative and true positive rates.

The paper enhances metabolite annotation confidence by integrating taxonomic information into spectral matching scores, improving recall and precision of annotation tools.

The paper identifies alpha-1 antitrypsin as a key contributor to GlycA's predictive power for disease risk and mortality across multiple conditions.

The paper introduces SF-Matching, a machine learning method for metabolite identification from MS/MS spectra using structural features, achieving high accuracy when combined with CSI:FingerID.

The paper introduces ViMMS, a simulator for LC-MS/MS fragmentation strategies, enabling in-silico testing and optimization of mass spectrometry acquisition methods.

WiPP improves GC-MS peak detection by combining multi-algorithm results with machine learning-based quality assessment, enhancing accuracy and enabling algorithm comparison.

The paper compares a GC-Orbitrap-MS with FID capabilities to traditional GC-MS for metabolomics of human serum, showing improved quantification and pathway coverage.

The paper generates and validates an untargeted lipidomics dataset from maternal plasma samples associated with preterm birth, providing a resource for further research.

The paper evaluates three NMR analysis tools for metabolite identification in human urine, finding ISA and ACP robustly identify metabolites while PCA fails.

The paper analyzes targeted metabolomics data from diverse biological samples to assess measurement quality and identify metabolite dynamics across conditions.

This work proposes algorithms to correct variations between mass spectra, improving comparability across large datasets and enabling machine learning applications.

The paper provides a metabolomics dataset from 30 mouse gene knockouts and wildtypes, identifying 832 metabolites across multiple analytical methods.

This paper introduces CIVM-NMR, a method for continuous in vivo metabolomics monitoring, and demonstrates its application in human cells and Neurospora crassa.

ASICS provides an R package for a complete workflow analyzing 1D 1H NMR spectra, improving metabolite quantification and identifying relevant metabolites in metabolomics studies.

The paper introduces a recursive algorithm and webserver, MetDNA, for large-scale and unambiguous metabolite identification in untargeted metabolomics.

MASST enables querying of public small-molecule tandem MS data to discover molecular relationships in metabolomics and natural product datasets.

This paper introduces a tool that uses expert knowledge in machine learning to analyze dynamic metabolomics data, showing it performs as well as statistical methods.

The paper develops a machine learning method to classify molecular formula assignments into lipid categories using SMIRFE data, achieving high accuracy and precision.