Department Environmental Chemistry

Analytic software

MLinvitroTox

MLinvitroTox is an open-source Python package for fully automated high-through-put hazard-driven prioritization of unknown high-resolution mass spectrometric signals in complex environmental samples. It is a machine-learning based system comprising 490 independent XGBoost classifiers trained on molecular fingerprints of chemical structures and target-specific endpoints from the ToxCast/Tox21 invitroDBv4.1 database. MLinvitroTox predicts a bioactivity fingerprint for each unidentified HRMS signal (with a specific m/z) based on the molecular fingerprints derived from MS2 fragmentation spectra. The binary 490-bit bioactivity fingerprints are used to prioritize the HRMS features for further identification.

envihomolog

Web-interface for the detection and visualization of LC-MS signal patterns caused by homologous compounds, based on the algorithm from the R nontarget package. Requires upload of picked and preferably deisotoped LC-MS peak lists. Results can be downloaded as .csv-files. Best viewed with Firefox, Google Chrome or Safari.

enviPat website

Based on the R package of the same name, web interface for calculating isotope patterns (fine structure), profiles, centroids and intensoids. Resolves isotopic fine structures even for large molecules. Allows usage of build-in instrument resolutions, predefined or user-defined adducts and batch processing. Best viewed with Firefox, Google Chrome and Safari. 

R package 'nontarget'

R tool for grouping of isotopes and adducts to components as well as detection of homologue series out of high resolution mass spectrometer data sets.

R package 'RMassBank'

R tool for batch-processing, recalibration and cleanup of HRMS spectra and MassBank record generation.

RMassBank is a workflow to enable the creation of high quality accurate mass MS/MS spectral records for the MassBank database. The workflow includes automatic recalibration and processing of mass spectra within R from mzML files and the retrieval of compound information via web services to minimise the user efforts in record generation. At this stage RMassBank is designed and tested on Orbitrap XL (Thermo) data. For more information please contact Johannes Boog.

Example data for RMassBank (raw data files, ~50 MB)

R package enviPick

Partitioning, EIC clustering and peak detection for baseline-corrected and centroided LC-HRMS data in mzXML format. Interactive result and raw data viewer, batch processing & browser-based UI.

Simulation Software

Assessment tool for OECD 308 experiments

A command-line utility for the inverse modelling of OECD 308 experiments with Bayesian parameter inference.