MS-DIAL

Data processing pipeline for untargeted lipidomics

Description

MS-DIAL is an open source software for untargeted metabolomics and lipidomics data. The current version 4 series supports GC-MS, GC-MS/MS, LC-MS (Full MS), LC-DDA, LC-DIA (SWATH & all ion fragmentation), LC-Ion mobility (IM)-DDA, and LC-IM-DIA. Lipid annotation is based on retention time (optional), CCS (optional), m/z, and MS/MS (recommended) using a hybrid scoring system of classical spectral matching algorithm and defined fragmentation rules for each lipid subclass. The program provides an all-in-one solution from data import of MS raw data until lipidome table export (such as mztab-M) and statistical analyses.

Technical Information

Publications:

PMID:32541957

Training datasets:

Yes

Documentation and user guide:

Mtbinfo ,

Metabolomics

Download / Web-service link:

Yes

Programming languages:

Platforms:

Windows,

Linux,

MacOS

Output formats:

TSV

Input formats:

mzML,

netCDF,

abf,

ibf,

.d(Agilent),

.d(Bruker),

.wiff(SCIEX),

.wiff2(SCIEX),

.raw(Thermo),

.raw(Waters)

Web platform:

Desktop client:

Yes

CLI:

Yes

GUI:

Yes

License:

LGPL v3 (code) / CC-BY-SA (software)

Tasks

6) Analysis and visualization of lipidomics data

Direct connection to identification results:

Yes

Classification and feature selection:

PLS-DA,

OPLS-DA

Clustering and correlation analysis methods:

HCA

Feature identification:

Supervised multivariate statistical analysis methods:

PLS,

PLS-DA,

O-PLS,

O-PLS-DA

Unsupervised multivariate statistical analysis methods:

PCA

Univariate statistical analysis methods:

ANOVA,

Fold-change analysis

Missing data handling:

Yes

Data pre-treatments:

Filtering by blank’s peaks,

An internal standard based (for relative intensity-based sample comparison),

Quality control sample’s profiles based (for LOWESS method),

Total ion count based (for relative intensity-based sample comparison),

Scaling

4.1) Full MS (HRAM LC-MS)

Other features:

See Lipid identification using DDA dataset (Data dependent acquisition 4.2)

Batch processing:

Yes

Lipid identification:

Build-in matching to the MS-DIAL internal database or user-defined library.

Correction for in source fragments:

Yes

Support multiple adducts:

Yes

Algorithms:

MS1 accuracy matching to a library (the MS-DIAL internal database or user-defined library)

LC-MS:

Yes

Shotgun MS/MS:

4.2) Data dependent acquisition (DDA)

Other features:

Isotopic deconvolution,

Adduct/cluster/in-source ion grouping,

Alignment

Batch processing:

Yes

Correction for in source fragments:

Yes

Support multiple adducts:

Yes

Spectra annotation:

Yes

Scores:

Global scoring that takes into account retention time, mass accuracy, isotopic pattern and MS2 fragmentation

Algorithms:

MS1 accuracy, MS2 library matching, rule-based fragment assignment

Requirements:

User can load user-defined library (not obligatory, the internal library is included in the MS-DIAL environment)

Decision rules:

Yes

Link to spectra matching:

Yes

Spectra matching:

Spectra libraries are included in the MS-DIAL environment, and the library is also available as MSP.

LC-MS:

Yes

Shotgun MS/MS:

4.3) Data independent acquisition (DIA)

Other features:

See Lipid identification using DDA dataset (Data dependent acquisition 4.2)

Additional requirements:

The dictionary file including the precursor isolation window information is needed

MS-MS2 Deconvolution algorithm:

https://www.nature.com/articles/nmeth.3393 ,

MSDec algorithm, see

SWATH like:

Yes

AIF like:

Yes

Input formats:

abf,

.d(Agilent),

.d(Bruker),

.wiff(SCIEX),

.wiff2(SCIEX),

.raw(Thermo),

.raw(Waters)

4.4) Tools considering ion mobility separation

Other features:

See Lipid identification using DDA dataset (Data dependent acquisition 4.2)

Annotate CSS values to lipids:

Yes

Use CSS for identification:

https://www.nature.com/articles/nmeth.3393 ,

The CCS value is used for scoring and filtering candidates, the score is calculated by the equation following gaussian distribution, see

Accepted data formats:

IBF,

.d(Agilent),

.d(Bruker),

.raw(Waters)

4.5) Identification of oxidized lipids

Other features:

Isotopic deconvolution, adduct/cluster/in-source ion grouping.

Spectra annotation:

Yes

Scores:

Global scoring that takes into account mass accuracy, isotopic patter and MS2 fragmentation

Algorithms:

MS1 accuracy, MS2 library matching

Requirements:

http://prime.psc.riken.jp/compms/msdial/main.html ,

Oxidized lipids spectra library (build-in, and it is also available as MSP format) see

Decision rules:

Yes

Spectra matching:

Yes

Lipid coverage:

GPL

5) Lipid quantification from untargeted lipidomics datasets (HRAM MS, DDA, DIA)

Calibration curve:

Yes

Semi absolute quantification using internal standards:

SPLASH Lipidomics I,

SPLASH Lipidomics II,

EquiSPLASH

Relative quantification:

Yes

Normalization methods:

Internal standard based (for relative intensity-based sample comparison),

Quality control sample’s profiles based (for LOWESS method),

Total ion count based (for relative intensity-based sample comparison)

Quantification by:

Peak area

Adjustable feature detection:

Yes

Peak processings:

Peak picking,

Peak alignment,

Deconvolution