mzmine: Free LC-MS Data Processing Workflow

mzmine is open-source software that turns raw mass spectrometry data (LC-MS, GC-MS, ion mobility, direct infusion and MS imaging) into an aligned, annotated feature table. This guide is for researchers who have raw files and need to know what mzmine does, how to install it, how its processing wizard runs from import to batch, and where the output goes next. Facts were read from the mzmine project pages and documentation on 2026-11-08.

What mzmine is used for in research

mzmine is a data-processing program for mass spectrometry. According to the project site (mzio.io), it processes raw data from LC-MS, GC-MS, ion mobility, direct infusion and MS imaging, runs on your own hardware, and has been open source since 2005. The usual end product is an aligned feature table: one row per detected feature (an m/z and retention-time signal) with its intensity in every sample, plus annotations where they could be made.

The documented processing chain is:

Import of data from the major instrument vendors and open formats such as mzML.
Mass detection, including centroiding and noise filtering.
Feature resolving: chromatogram building and peak resolving.
Alignment of features across samples into one feature table.
Ion identity networking, which groups adducts, isotopes and in-source fragments that belong to the same compound.
Annotation by spectral library matching and in-silico prediction.

For the instrument side of this workflow, see LC-MS explained and choosing a metabolomics MS platform and QC samples.

Licence and editions

The mzmine source code is distributed under the MIT license (GitHub repository, read 2026-11-08). The project describes two editions: mzmine COMMUNITY, free for academic use, and mzmine PRO for commercial use, which adds licensing, support and IT integration. Check the project site for current terms before using it in a company or core-facility setting.

Install or access (official source only)

Download an installer or a portable version from the official releases page: github.com/mzmine/mzmine/releases/latest. Installers exist for Windows, macOS and Linux.
You do not need to install Java. The GitHub readme states that mzmine packages a specific Java Virtual Machine.
On Debian or Ubuntu-style Linux, download the .deb file and run sudo apt install mzmine*.deb. To check the install, run /opt/mzmine/bin/mzmine -help. Optional dependencies listed by the project are xdg-utils, libgl1, libgtk-3-0 and libxtst6.
Building from source (./gradlew) is for developers and needs JDK 23 or newer; end users should use the packaged release.

Step-by-step: process a dataset with the processing wizard

The documentation’s processing wizard builds a batch for you. Its pages run in this order (mzmine documentation, “Processing wizard”, read 2026-11-08):

Data import: drag and drop your raw files.
Sample introduction system: choose how the sample entered the instrument, for example MALDI, HPLC or DESI.
Ion mobility instrument: set this if your data has an ion mobility dimension.
MS instrument: choose the mass detector and set noise thresholds, minimum feature height and m/z tolerances.
Filters: choose whether to keep or remove the original feature lists, the minimum number of samples per aligned feature, and an optional 13C isotope pattern filter.
Annotation: add a local compound database (csv or tsv), optional lipid annotation, and spectral library files.
Workflows: choose one of DDA, DIA, Library generation, Imaging, Target plate or Spectral deconvolution, then set the export path and a base file name. The wizard creates a new folder, and each export module adds its own suffix to the base name. An Advanced mode switch reveals a Parameter Customization tab.
Create batch: the Create batch button opens the batch queue; click OK to start processing.

In the batch queue you can double-click a step to modify it. Overrides are made through Select Module, Parameters and the Parameter Editor, applied to the first, last or all occurrences of a module, and listed in a Parameter Overrides table (Override, Remove, Clear All). The Estimate and Optimize functions reset any customization. For repeatable or server runs, the documentation also describes a command-line tool for batch mode.

Output and how to read it

The main result is the aligned feature list shown inside mzmine, which you can export from the Feature list methods menu. Two documented exports connect to downstream tools (documentation search listing, read 2026-11-08; menu wording may differ between versions):

Export	Goes to	What it is for
GNPS – feature based molecular networking	GNPS	Molecular networking and library matching of MS/MS spectra; the docs recommend the preset “Single scan: Merged across energies”.
MetaboAnalyst export	MetaboAnalyst	A csv with ID, m/z, retention time and preferred identity name, for MetaboAnalyst’s statistical analysis.

Read the feature table as a measurement matrix, not a list of identified compounds: a feature is a detected signal, and only features with an annotation have a proposed identity. Annotation by library matching or in-silico prediction proposes an identity and is not by itself confirmation.

Troubleshooting

These entries come from the official mzmine troubleshooting page (read 2026-11-08).

Problem	Documented cause and fix
mzmine does not start	The configuration may be corrupted. Delete the `.mzmine.conf` file in your user folder (it may be hidden) and reinstall mzmine.
mzML files do not import	The file path written into the mzML file during conversion contains characters mzmine cannot parse. Reconvert with input and output directories whose paths use only Roman characters.
Thermo raw files do not import (“No scans found”)	Delete the `mzmine_thermo_raw_parser` folder from both the mzmine temp folder and the system temp folder.
Scans not sorted in retention time (often Waters data)	Reconvert the files with the Waters data connect tool.
“Scans were recognised as empty” warning	Not an error. If conversion used a threshold, consider lowering it; with no thresholding it is usually a false positive, as it often is with vendor formats.
Zero intensity values in scans	Can occur with an old MSConvert version when the peak picking filter runs after the title maker. Reconvert, ideally with a current MSConvert. This is only a warning.
GUI freezes on a remote desktop after logging out and back in	A known issue with no documented fix; the page says it may be fixed in future JavaFX versions.

Free alternatives and neighbouring tools

GNPS and MetaboAnalyst are not replacements for mzmine: they sit after it. mzmine builds the feature table from raw files; those tools analyse and network it. The table shows how the three relate.

Tool	Role relative to mzmine	CASRAI guide
mzmine	Raw data to aligned, annotated feature table	This page
GNPS	Receives an mzmine export for feature-based molecular networking	GNPS guide
MetaboAnalyst	Receives an mzmine export for statistical analysis	MetaboAnalyst guide

FAQ

Is mzmine free?

The source code is MIT-licensed, and the project describes mzmine COMMUNITY as free for academic use and mzmine PRO as the edition for commercial use (project site, read 2026-11-08).