LipidSig: a web-based tool for lipidomic data analysis

Lipidomics technology provides a fast and high-throughput screening to identify thousands of lipid species in cells, tissues or other biological samples and has been broadly used in several areas of studies. In this page, we present an overview that gathers comprehensive analyses that allow researchers to explore the quality and the clustering of samples, correlation between lipids and samples, and the expression and composition of lipids.

Data Source

Demo dataset

Adipose tissue ATGL modifies the cardiac lipidome in pressure-overload-induced left Ventricular failure (PLoS Genet. 2018)

Lipid expression data can be uploaded by users or example datasets are also provided. The Lipid expression data , and Lipid characteristics (optional), needs to be uploaded in CSV format.

Lipid characteristics

: Successfully uploaded. : Error happaned. Please check your dataset. : Warning message.

Lipid expression data

Processed data
Uploaded data

Lipid characteristics

Result

Cross-sample variability

In this page, three types of distribution plot provide a simple view of sample variability. The first histogram depicts the numbers of lipids expressed in each sample. The second histogram illustrates the total amount of lipid in each sample. The last density plot visualizes the underlying probability distribution of the lipid expression in each sample (line). Through these plots, users can easily compare the amount/expression difference of lipid between samples (i.e., patients vs. control).

Dimensionality reduction

Dimensionality reduction is common when dealing with large numbers of observations and/or large numbers of variables in lipids analysis. It transforms data from a high-dimensional space into a low-dimensional space so that to retain vital properties of the original data and close to its intrinsic dimension. Three dimensionality reduction methods are provided in this page, PCA, t-SNE, UMAP.

Dimensionality reduction method:

Scaling:

Centering:

NOTE: Scaling and Centering are on by default. Centering is strongly recommended for pre-processing steps.

PCA:

Perplexity:

Number of iterations:

Scaling:

Number of neighbors:

Distance metric:

Clustering method:

Number of groups:

Distance metrics:

Number of groups:

Distance measure:

Clustering method:

Epsilon:

minPts:

Correlation heatmap

Correlation heatmaps illustrate the correlation between lipid samples or characteristics and depict the patterns in each group. The correlation can be calculated by Pearson or Spearman. The correlation coefficient is clustered depending on the user-defined method and distance. Furthermore, users have to select a lipid characteristic to display the heatmap. Two heatmaps will be shown by lipid samples or by characteristics.

Correlation method:

Distance measure:

Clustering method:

By samples

Download matrix

By lipid characteristics

Download matrix

Lipid characteristics profiling

In this page, users can discover lipid expression over specific lipid characteristics by scrolling dropdown menu. Lipids will be firstly classified by the selected characteristics from ‘Lipid characteristics’ table uploaded by users. Next, the lipid expression will be shown in bar plot, which depicts the expression level of each sample within each group (e.g., PE, PC) of selected characteristics (e.g., class). Additionally, a stacked horizontal bar chart reveals the percentage of characteristics in each sample. For instance, if users select class as lipid characteristics from the dropdown menu, the stacked bar chart will tell users the percentage of TAG, ST, SM etc. of each sample, the variability of percentage between samples can also be obtained from this plot.

Select the lipid characteristic: