Welcome to LipidAnalyst!
LipidAnalyst: A Comprehensive Tool for Lipidomics Data Analysis
LipidAnalyst is an interactive web application designed for the analysis of lipidomics data. It provides a user-friendly interface for uploading, processing, normalizing, and analyzing lipidomics datasets.
For detailed instructions and help, please refer to our tutorial.
TutorialFor support or inquiries, please contact: lipidanalyst-requests@umich.edu
What's New
July 2026
LipidAnalyst is now published!
Our LipidAnalyst paper has been published in Metabolites .
Read the paperIf you use LipidAnalyst in your research, please cite our paper:
Liu, X., et al. LipidAnalyst: A Comprehensive Tool for Lipidomic Data Visualization and Analysis. Metabolites. (2026).
Recent Updates
- Expanded lipid annotation.
- Improved LipidAnalyst documentation and tutorial.
Upload Lipidomics Data
Please select the correct data format before uploading the file. After adjusting any settings, make sure to reload your data so that the software uses the correct parameters.
Example data 1: Targeted lipidomics of plasma from sham and CKD mice following 16 weeks of high-fat diet (HFD).
Saum, K., et al. Chronic kidney disease induces distinct alterations of macrophage lipid metabolism in a mouse model of atherosclerosis. Journal of Lipid Research. (2026). View publication
Download Example Data 2Example data 2: MS-based shotgun lipidomics analysis of human plasma samples from HFrEF-patients (n = 13) and non-HFrEF controls (n = 10).
Salatzki, J., et al. Adipose tissue ATGL modifies the cardiac lipidome in pressure-overload-induced left ventricular failure. PLoS Genet. (2018). View publication
Lipidomics Data Preview
Upload Metadata (Group information)
Metadata Preview
Upload Internal Standard
Please select the correct data format before uploading the file. After adjusting any settings, make sure to reload your data so that the software uses the correct parameters.
Please select the correct data format before uploading the file.
Internal standard information is not available for example dataset 2. Please skip this step if you use example dataset 2.
Internal Standard Preview
Data Filtering Overview
Data filtering is a crucial preprocessing step in lipidomics data analysis. It helps to eliminate noise and irrelevant features, ensuring that subsequent analyses focus on meaningful biological variations.
This section provides options to apply various filters to your lipidomics data. You can choose to enable or disable each filter and set specific thresholds according to your analysis needs.
Low Quality Filter
Low Quality Filter removes features with a high percentage of missing values across samples.
Low Abundance Filter
Low Abundance Filter removes features with a mean lower than a specified percentile across all feature means.
Low Variance Filter
Low Variance Filter removes features with low variability across samples, which would not be useful for distinguishing between different conditions or groups.
Filtered Data Preview
Missing Value Imputation Guide
Missing values are common in lipidomics datasets and can arise from several sources, including:
- Technical limitations of the mass spectrometer, such as detection thresholds or signal suppression,
- Variability introduced during sample extraction, handling, or instrument runs,
- True biological absence of a lipid in certain samples or conditions.
Understanding the pattern of missingness is essential for selecting an appropriate imputation strategy. Different types of missingness can reflect different underlying causes and may require tailored handling to avoid introducing bias into downstream analyses such as statistical testing, or network construction.
We categorize missingness into two major types:
- Group-level missingness: A lipid is almost completely missing within one experimental group (e.g., all disease group samples have NA). This often suggests true biological absence or very low abundance. In these cases, methods such as Limit of Detection (LoD) imputation is generally more appropriate.
- General missingness: Values are sporadically missing across samples but not confined to a single group. This pattern typically reflects technical noise or stochastic signal dropout. For this scenario, K-Nearest Neighbors (KNN) imputation(sample-wise) is recommended as it leverages similarity among samples to estimate reasonable values.
You can view the Missing value heatmap to identify these patterns and decide on the best imputation approach for your data.
Group-level Missingness Summary
Group Level Missing means features that are missing in a large proportion of samples within a specific group. This type of missingness can usually mean features are truly absent in the specific group, and we suggest to use Limit of Detection (LoD) 1/5 minimum value of the specific feature to impute the missingness here.
Imputation Settings
This panel allows you to handle general missing values in your lipidomics data. You can choose to skip imputation (if no missingness) or select from various imputation methods to fill in missing data points.
Imputed Data Preview
Combine duplicated lipids or lipids with different adducts
In lipidomics data, it is common to encounter duplicated lipid entries that represent the lipid species with same shorthand name but differ in their adduct forms (e.g., [M+H]+, [M+Na]+, [M+K]+). Combining these duplicated lipids into a single representative entry can help streamline data analysis and interpretation. This process involves selecting criteria to identify and merge these duplicates based on their lipid names and adduct types.
Combine Duplicated lipids
Combined Data Preview
Lipid Parsing Controls
Chain annotation convention
The biological meaning of each chain field depends on lipid class:
- Sphingolipids (e.g., SM, Cer, dhCer, LacCer, HexCer):
Chain1= sphingoid base (long-chain base, LCB),Chain2= N-acyl fatty acid (amide-linked acyl chain),Chain3= not applicable. - Phospholipids (e.g., PC, PE, PG, PI, PS,LPC, LPE, PE(O), PE(P), PA):
Chain1andChain2refer to fatty acyl chains at Sn1 and Sn2 carbons of the glycerol back bone, respectively.Chain3= not applicable. - Glycerolipids (e.g., MAG, TAG, DAG):
Chain1toChain3refer to fatty acyl chains at Sn1 to Sn3 carbons of the glycerol back bone, respectively. For MG,chain2andchain3are not applicable. - Cholesteryl esters (CE):
Chain1= esterified fatty acid chain,Chain2andChain3= not applicable.
The parsing table is editable. Double-click a cell to make changes. If any lipid information is incorrect, you may revise it directly in the table.
Use the search box in each column to filter the rows that you want to view or edit.
If you want to edit multiple points in the parse table,
we suggest download the parse table as a CSV file, make changes in Excel,
and upload the updated parse table using the file input below.
Parsed Lipid Names
Lipid Class Pie Plot
Lipid Class Boxplot
PCA Plot
Sample Boxplot
Barplot of lipid species
Barplot of group distribution
Internal Standard Selection
Normalization by internal standards is a widely used technique in lipidomics data analysis to correct for technical variability and improve the accuracy of quantification. Internal standards are compounds that are chemically similar to the target analytes but are not naturally present in the sample. By adding known quantities of internal standards to each sample prior to analysis, researchers can account for variations in sample preparation, instrument performance, and other factors that may affect the measured signal intensity.
Updated Parsed Table with Internal Standard
Normalization Messages
Normalized Data Preview
Normalization by a user-defined constant value
Normalization by a constant value involves adjusting the lipid abundances by dividing each data entry by a user-defined constant.
This method is useful when there is a known factor that affects all samples equally, such as dilution factor, weight, or volume.
By applying this normalization, researchers can standardize the lipid measurements across samples, facilitating more accurate comparisons and analyses.
Metadata Supplement
Normalization by metadata involves adjusting lipid abundances based on relevant metadata information.
This method helps to account for variations in sample characteristics, experimental conditions, or other factors that may influence lipid measurements, such as cell counts and protein concentration.
By normalizing the lipid data using appropriate metadata, researchers can improve the accuracy and reliability of their analyses, leading to more meaningful biological interpretations.
Normalized Data Preview
Normalization
Data Transformation
Data Scaling
Normalized Data Preview
Boxplot Settings
Pixel size settings for downloading Boxplot
Font Settings
Boxplot Output
Plot Settings
Pixel size settings for downloading plots
Font and Color Settings
Plot Settings
Pixel size settings for downloading plots
Font and Color Settings
Plot Output
Differential Mean Lipid Heatmap Parameter
Pixel size settings for downloading Heatmap
Font and Color Settings
Differential Mean Lipid Heatmap Output
Volcano Plot Settings
Volcano plots are a type of scatter plot that are used to visualize the results of differential expression analyses. They plot the negative logarithm of the p-value on the y-axis against the log fold change on the x-axis, allowing for easy identification of features that are both statistically significant and have a large magnitude of change.
Features significancy is determined by welch's t-test and adjusted p values (Benjamini-Hochberg method).
Difference was defined as the difference in mean lipid abundance between the two groups. This metric was used instead of a ratio because normalized or transformed abundance values may include negative values, making ratio-based fold change inappropriate.
Pixel size settings for downloading plots
Font and Color Settings
Volcano Plot Output
Volcano Plot Data
PCA Plot Settings
Pixel size settings for downloading PCA
Font and Color Settings
PCA Plot Output
Heatmap Settings
Heatmap Output
Lipidomics Mean Calculator
This tool calculates the mean and standard deviation of lipid abundances for each lipid feature within each sample group.
It provides a summary of the average lipid levels and their variability within each group, facilitating comparisons between different experimental conditions or sample groups.
Mean Calculated Data Preview
ANOVA Settings
ANOVA (Analysis of Variance) is a statistical method used to compare the means of three or more groups to determine if there are significant differences among them. It assesses whether the observed variations in the data can be attributed to the group factor or if they are likely due to random chance.
ANOVA Results
ANOVA PCA Plot
Correlation
Correlation heatmap
DSPC Network Settings
DSPC (Debiased Sparse Partial Correlation) is a statistical framework designed to infer sparse and interpretable molecular networks from high-dimensional omics data such as lipidomics. It estimates partial correlations while correcting for the bias introduced by regularization (e.g., graphical lasso), allowing more reliable detection of direct molecular relationships.
DSPC Network Output
PLS-DA Model Overview
Partial Least Squares Discriminant Analysis (PLS-DA) is a supervised multivariate statistical method used to model the relationship between lipidomic data and predefined categorical groups.
PLS-DA identifies latent variables that maximize the covariance between the lipid features and the group labels, enabling effective classification and discrimination of samples based on their lipid profiles.
This technique is particularly useful for biomarker discovery, as it highlights the most relevant features contributing to group separation while handling complex, high-dimensional datasets typical in lipidomics studies.
PLS-DA Model Settings
PLS-DA Model Output
OPLS-DA Model Overview
Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA) is a supervised multivariate statistical method used to identify differences between predefined groups in complex datasets.
Unlike PLS-DA, OPLS-DA explicitly separates the variation in the data into:
- Predictive components: directly associated with the group labels.
- Orthogonal components: capture systematic variation unrelated to group differences (e.g., technical variation or biological heterogeneity).
OPLS-DA Model Settings
OPLS-DA Model Output
Random Forest Model Settings
Random Forest is an ensemble machine learning algorithm that builds multiple decision trees and combines their predictions to improve classification accuracy.
It is well suited for high-dimensional lipidomics data and provides feature importance rankings.
Random Forest Model Output
Download All Generated Files
Thank you for using our lipidomics data analysis tool! If you have any feedback or encounter any issues, please don't hesitate to contact us.
For support, please contact: lipidanalyst-requests@umich.edu
Click the button below to download all output files as a zip archive.
Download All Data (ZIP) Download All Plots (ZIP)