ssGSEA outputs¶
This page describes the single-sample gene set enrichment outputs generated by MTD Explorer.
ssGSEA summarizes gene set activity at the sample level.
In MTD Explorer, these outputs help inspect whether host gene-set activity differs across biological groups.
The main folder is:
Main ssGSEA figures¶
A typical ssGSEA/ folder may contain:
plots_GO_names_ssGSEA_PCA_samples.png
plots_GO_names_ssGSEA_sample_correlation_heatmap.png
plots_GO_names_ssGSEA_top_differential_boxplots.png
plots_GO_names_ssGSEA_top_variable_heatmap.png
plots_ssGSEA_PCA_samples.png
plots_ssGSEA_sample_correlation_heatmap.png
plots_ssGSEA_top_differential_boxplots.png
plots_ssGSEA_top_variable_heatmap.png
The two PCA figures can be visually identical in some runs.
For documentation purposes, this page shows one representative PCA figure.
ssGSEA PCA¶
The representative PCA figure is usually:
This PCA summarizes global variation in sample-level gene-set activity.
Samples that cluster close together have more similar ssGSEA profiles.
Samples that separate strongly along the main principal components have larger global gene-set activity differences.
PCA is exploratory and should be interpreted together with sample metadata, gene-set definitions, and the underlying ssGSEA score matrix.
GO-name ssGSEA sample correlation heatmap¶
The GO-name sample correlation heatmap is usually:
This heatmap shows similarity among samples based on their ssGSEA profiles.
It is useful for checking whether samples from the same group have similar gene-set activity patterns.
GO-name ssGSEA top differential boxplots¶
The GO-name top differential boxplot figure is usually:
This figure summarizes selected gene sets with stronger group-level differences.
It is useful for quickly identifying which gene sets may drive differences between biological groups.
Use the corresponding tables for statistical interpretation.
GO-name ssGSEA top variable heatmap¶
The GO-name top variable heatmap is usually:
This heatmap shows the most variable GO-named ssGSEA features across samples.
It helps identify gene sets with strong sample-to-sample variation.
ssGSEA sample correlation heatmap¶
The sample correlation heatmap is usually:
This figure provides an alternative sample-level correlation view using the ssGSEA feature labels generated by the pipeline.
ssGSEA top differential boxplots¶
The top differential boxplot figure is usually:
This figure highlights selected ssGSEA features that differ between groups.
It is useful as a visual summary, but statistical conclusions should be based on the output tables.
ssGSEA top variable heatmap¶
The top variable heatmap is usually:
This heatmap summarizes the most variable ssGSEA features across samples.
It can help identify gene-set activity patterns that separate samples or reveal possible outliers.
Recommended inspection order¶
For ssGSEA outputs, inspect:
ssGSEA/plots_GO_names_ssGSEA_PCA_samples.png
ssGSEA/plots_GO_names_ssGSEA_sample_correlation_heatmap.png
ssGSEA/plots_GO_names_ssGSEA_top_variable_heatmap.png
ssGSEA/plots_GO_names_ssGSEA_top_differential_boxplots.png
ssGSEA/plots_ssGSEA_sample_correlation_heatmap.png
ssGSEA/plots_ssGSEA_top_variable_heatmap.png
ssGSEA/plots_ssGSEA_top_differential_boxplots.png
methods/mtd_methods_run_parameters.csv
The methods/mtd_methods_run_parameters.csv file records run settings and
software versions.
What these outputs can support¶
ssGSEA outputs can help answer whether host gene-set activity profiles cluster by group.
They can also help identify gene sets with strong variation across samples or large group-level differences.
What not to conclude¶
Do not interpret PCA separation alone as proof of differential gene-set activity.
Do not interpret heatmap clustering without checking metadata and the underlying ssGSEA score matrix.
Do not interpret a boxplot as final statistical evidence without checking the corresponding tables and the study design.
When outputs may be missing¶
ssGSEA outputs may be missing when host expression matrices are unavailable, gene identifiers cannot be mapped to gene sets, too few samples are available, or the ssGSEA score matrix is too sparse for plotting.