Florian Huber

Beyond the Default: Optimizing Molecular Networking with arteMIS

The paper introduces arteMIS, a framework for optimizing molecular network parameters in metabolomics, showing improved network quality over default settings.

Count your bits: fingerprint benchmarking to assess broad chemical space representation

The paper benchmarks molecular fingerprints across multiple datasets and tasks, finding that count and unfolded variants improve specificity and alignment with structural references.

Bridging polarities in metabolomics: Cross-ionization mode chemical similarity prediction between tandem mass spectra

MS2DeepScore 2.0 predicts chemical similarity between tandem mass spectra across ionization modes, improving metabolite annotation and data exploration in untargeted metabolomics.

Mass2SMILES: deep learning based fast prediction of structures and functional groups directly from high-resolution MS/MS spectra.

Mass2SMILES predicts chemical structures and functional groups from MS/MS spectra using a deep neural network, achieving high accuracy and speed for metabolomics data.

Unified and standardized mass spectrometry data processing in Python using spectrum_utils

The paper updates the spectrum_utils Python package to integrate mass spectrometry standards, enhance data processing, and unify visualization across proteomics and metabolomics applications.

Comparison of Cosine, Modified Cosine, and Neutral Loss Based Spectrum Alignment For Discovery of Structurally Related Molecules

The paper compares cosine, modified cosine, and neutral loss methods for MS/MS spectrum alignment to find structurally related molecules, finding modified cosine performs best.

MS2Query: Reliable and Scalable MS2 Mass Spectral-based Analogue Search

MS2Query improves metabolite annotation by combining mass spectral embeddings and precursor masses to rank analogues and exact matches more reliably and scalably.

MEMO: Mass Spectrometry-based Sample Vectorization to Explore Chemodiverse Datasets

MEMO enables clustering of chemodiverse samples using LC-MS/MS profiles without feature alignment, achieving performance comparable to state-of-the-art methods faster.

MS2DeepScore - a novel deep learning similarity measure for mass fragmentation spectrum comparisons

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

Spec2Vec: Improved mass spectral similarity scoring through learning of structural relationships

Spec2Vec improves spectral similarity scoring by learning fragmental relationships, showing better correlation with structural similarity than cosine-based scores in metabolomics analyses.

matchms - processing and similarity evaluation of mass spectrometry data

The paper introduces matchms, a tool for processing and evaluating similarity in mass spectrometry data, addressing challenges in dataset consistency and metadata quality.