Lipids are important components of cell membranes and lipid particles (e.g., lipoproteins) and play many important roles in cellular functions, including cellular barriers, membrane matrix, and signaling. Abnormalities in their metabolism may trigger many human diseases, such as atherosclerosis. Lipidomics is an emerging discipline that comprehensively and systematically analyzes and studies lipids and their interacting molecules in organisms, cells, or tissues to understand the composition, structure, and function of lipids, and then reveals the relationship between lipid metabolism and the physiopathological processes of the organism. Currently, lipidomics has been widely used in biomedical research, drug development and other important fields.
To facilitate lipidomic research, numerous databases and software tools have been developed to store, analyze, and visualize lipidomic data. Here is a brief description of a few common ones.
LIPID MAPS
LIPID MAPS, an initiative by the National Institutes of Health (NIH), serves as a comprehensive platform for lipidomics research. It offers access to a vast repository containing over 40,000 lipid compounds, along with authoritative classifications, nomenclature, and structural information.
This platform provides diverse modules catering to various aspects of lipid analysis and exploration. Users can conduct bulk searches for mass spectrometry data, access tutorials for guidance, utilize lipidomics analysis software, and employ mass spectrometry tools for precise analysis.
Additionally, researchers can leverage lipid structure drawing tools for visual representation and statistical analysis tools for data interpretation. The core of LIPID MAPS lies in its structure database and classification system, complemented by a gene/protein database offering molecular insights into lipid metabolism.
Furthermore, the platform offers access to lipidomics experimental data, empowering researchers with real-world insights. It also provides bioinformatics tools for predictive analysis based on specific parameters, such as mass-to-charge ratio or tandem mass spectra.
LIPID MAPS category annotation and heatmap analysis of differential metabolites of polyketides (An Y et al. 2022).
LipidIMMS Analyzer
Ion mobility-mass spectrometry (IM-MS) exhibits tremendous potential in the field of lipidomics. However, lipidomics based on IM-MS is greatly hindered by the limitations of lipid structure identification software. The LipidIMMS Analyzer, an open-source software provided by the research group led by Zhu Zhengjiang in China, offers automated data processing and lipid identification, addressing various types of IM-MS-based spectral data.
This software represents a significant advancement by integrating a large-scale database covering over 260,000 lipid species, along with three-dimensional structural information for each lipid. This includes precise mass-to-charge ratio (m/z), retention time (RT), collision cross-section (CCS), and MS/MS spectra. The integration of multidimensional information facilitates lipid identification, significantly enhancing coverage and confidence in identification.
Users can simply import MS peak tables and MS/MS data files for analysis. The software supports different IM-MS instruments and data acquisition methods. Its user-friendly visual interface enables quick entry into lipidomics, making it accessible to researchers regardless of their level of expertise.
MS DIAL
MS-DIAL is a lipid data analysis software developed by the team led by Professor Oliver Fiehn from the University of California, Davis, and Professor Masanori Arita from RIKEN CSRS (and NIG). It integrates information from the LipidBlast lipid database and is an open-source tool. MS-DIAL offers 32 sub-classes with over 110,000 positive mode MS/MS spectra for lipid molecules and 48 sub-classes with over 150,000 negative mode MS/MS spectra.
This software supports mzML and major MS vendor formats, including Agilent (.D), AB Sciex (.Wiff), Thermo Fisher Scientific (.RAW), Bruker Daltonics (.D), and Waters (.RAW). Designed for large-scale sample analysis, MS-DIAL sequentially accesses raw data and retains peak information in memory, allowing for rapid processing times of around 1.2 minutes for a 600MB file.
Additionally, MS-DIAL is not only limited to lipidomics research but can also be used for the identification and analysis of small molecules in conventional metabolomics, including GC-MS data. Users can create their own spectral databases using this software, facilitating data analysis for large-scale sample sets.
LipidBank
LipidBank, established in 1989, originated from the Japanese Conference on Lipid Biochemistry (JCBL) as a publicly accessible natural lipid database. It encompasses a wide range of lipid categories, including fatty acids, glycerolipids, sphingolipids, steroids, and various vitamins. With over 6,000 unique molecular structures (in ChemDraw cdx and MDL MOL formats), LipidBank provides comprehensive lipid names (both common and IUPAC), spectral information (mass, UV, IR, NMR, etc.), and significant literature references.
This database exclusively focuses on natural lipid molecules, with all molecular information meticulously curated and approved by lipid research experts. LipidBank serves as a valuable resource for researchers and professionals in lipidomics, providing reliable data for exploration and analysis in the field of lipid biochemistry.
LipidHome
LipidHome is a centralized hub housing diverse computational lipidomics tools and resources. Researchers leverage LipidHome to access software applications for lipid structure prediction, lipidomics data analysis, and pathway modeling, thus streamlining lipidomic research workflows.
LipidFinder
LipidFinder automates lipid identification and quantification from liquid chromatography-mass spectrometry (LC-MS) data. Its advanced algorithms enable rapid and accurate analysis of complex lipidomic datasets, accelerating the pace of lipidomics research.
LipidHunter
LipidHunter offers a suite of features for lipidomic data analysis and interpretation. Researchers utilize LipidHunter for lipid identification, class analysis, and pathway mapping, leveraging its versatility to explore lipidomic datasets comprehensively.
LipidXplorer
LipidXplorer is a comprehensive software package facilitating lipidomic data processing and analysis. It integrates advanced algorithms for lipid identification, quantification, and statistical analysis, supporting targeted and untargeted lipidomics experiments with precision and efficiency.
LipidSuite
LipidSuite is an interactive web tool for differential and enrichment analysis of lipidomics. It provides a reliable workflow for targeted and untargeted lipidomic data analysis. Starting from data quality control, through data preprocessing, data exploration, difference analysis and enrichment analysis.
Users initiate the analysis by uploading two files containing lipidomic measurements and experimental sample annotations. LipidSuite accommodates various lipidomic data formats, including Skyline CSV exports, numeric matrices, and mwTab files.
The analysis begins with thorough data quality control to ensure data integrity and reliability. LipidSuite offers an "Overview" function, summarizing sample and lipid counts, along with lipid category distribution presented in interactive pie charts. Users can edit mismatched or incorrectly matched lipid names interactively.
Visual inspection of sample quality is facilitated through the "Samples" feature, providing interactive plots for quality assessment. Additionally, users can explore intensity distributions of specific lipid categories by selecting data subsets. The "Lipids" section enables visual inspection of lipid quality control plots to identify issues with individual molecules or categories, allowing for the removal of low-quality lipids or lipid categories if necessary.
Following data quality control, the preprocessing stage is executed. This includes summarization, imputation, and normalization of lipidomic data. Summarization is essential for targeted lipidomics to report multiple transitions of each lipid molecule accurately. LipidSuite offers various imputation methods, such as QRILC for low abundance data and k-nearest neighbors and singular value decomposition for random missing data. Post-imputation, histograms of imputation strength versus true measured values are displayed for evaluation. Finally, LipidSuite provides probability quotient normalization and internal standard normalization methods.
Data exploration is conducted through principal component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA) to visually explore the dataset.
For differential analysis, LipidSuite supports t-tests (two groups) and analysis of variance (ANOVA) for multiple groups. Upon completion, volcano plots and cluster heatmaps are generated to visualize differential analysis results. To assess lipidomic groups prioritized for change in selected comparisons, lipids are ranked by LogFC, and enrichment scores for each lipid group are calculated and visualized in enrichment analysis plots.
LipidSig
Lipids possess distinctive features, including lipid categories and chain lengths, setting them apart from typical metabolites. To streamline lipidomic analysis and enhance exploration of lipid characteristics, the LipidSig tool has been developed. This web-based tool offers a flexible and user-friendly interface, supporting data processing, mining, and visualization. It comprises five main sections: Profiling, Differential Expression, Correlation, Network, and Machine Learning, enabling assessment of lipid impacts on cellular or disease phenotypes.
The workflow of LipidSig consists of four key parts: (1) Data Upload, (2) Lipid Feature Conversion, (3) Data Processing, and (4) Functional Analysis and Visualization.
Lipid Feature Analysis evaluates changes in lipids classified based on one or multiple lipid features. LipidSig provides an automated conversion function, transforming lipid category expression into lipid features. Using this information, the original lipid expression data is transformed into new feature expressions. Users can select features to generate corresponding tables applied to analyses except for the Network section. Moreover, multi-feature analysis allows exploration of interactions among various features.
Profiling analysis enables researchers to effectively inspect data quality, sample clustering, correlation among lipid categories, and lipid feature composition. Differential Expression analysis integrates lipid-centric analyses to identify important lipid species/features. Correlation analysis aims to elucidate relationships between different clinical phenotypes and lipid features. Network analysis constructs lipid metabolism pathways based on the Reactome database and enriches networks with lipid-related gene queries. Machine Learning offers diverse feature selection methods and classifiers to build binary classification models. Subsequent analysis helps evaluate algorithm performance and explore important lipid-related variables.
Other Lipidomics Software
Additionally, there are numerous customized lipid databases and related software tools available, including Cyber Lipids, LIPIDAT, LipidHome, LIQUID, and more. Furthermore, major mass spectrometry instrument manufacturers have introduced their own commercial lipid software offerings. For instance, AB SCIEX's LipidView, combined with the complementary Lipidyzer product, enables qualitative and absolute quantitative lipid analysis. Thermo Fisher's LipidSearch contains a lipid database with over 1.5 million lipid ions and their predicted fragment ions, compatible with Thermo's triple quadrupole, ion trap, and Orbitrap LC-MS systems. Each of these offerings has its own unique features, allowing users to choose the most suitable product based on their specific needs.
Reference
- An, Yibo, et al. "The changes of microbial communities and key metabolites after early Bursaphelenchus xylophilus invasion of Pinus massoniana." Plants 11.21 (2022): 2849.