Biomarker Exploration (Gene Set Variation Analysis; GSVA)

The "Biomarker Exploration" module contains results of the Gene Set Variation Analysis (GSVA) algorithm applied to all samples within an experiment. Continue reading below to learn more about how these results can be interpreted.

What is Gene Set Variation Analysis (GSVA)?

Gene Set Variation Analysis (GSVA) is a method that uses gene expression data to estimate the "activity" of biological pathways in each individual sample. Instead of analyzing one gene at a time, GSVA looks at sets of genes that work together—such as in pathways related to immune response, metabolism, or cell signaling.

What Does Pathway "Activity” Mean?

In GSVA, "activity" refers to how strongly the genes in a pathway are collectively turned on or off in a sample.

More specifically:

A higher GSVA score indicates that the genes in the pathway are, as a group, more highly expressed relative to other genes in that sample. This suggests the pathway is more active, or “turned on.”
A lower GSVA score means those genes are less expressed, indicating the pathway is less active or “turned off.”

This score is relative within each sample, not absolute—it reflects the coordinated expression of the pathway’s genes compared to the overall gene expression profile of that sample.

How Is “Activity” Different From Normalized Expression?

It’s important to understand that pathway "activity" is not calculated simply as the average of normalized expression of all genes in the pathway.

Instead, GSVA evaluates:

How consistently and strongly the genes in the pathway are ranked among all genes in that sample.
Whether the genes in the set appear to be moving together (either up or down) in a coordinated way, suggesting a biological signal.
Enrichment of pathway genes at the high or low end of the expression spectrum within the sample.

This makes GSVA particularly powerful for detecting subtle but coordinated changes that might be missed by single-gene analyses.

How Does It Work?

GSVA uses a few core concepts to make the analysis intuitive and robust:

Unsupervised: GSVA doesn’t require pre-defined groups (e.g., case vs. control) to calculate scores. It evaluates each sample independently.
Non-parametric and rank-based: Instead of normalized expression values, GSVA uses the rank order of gene expression within each sample, making it more resilient to technical variability.
Sample-specific: It generates a pathway activity score for every pathway in every sample, which enables comparison of patterns across individuals, groups, or conditions.

This makes GSVA ideal for exploring patient heterogeneity, drug response, or tumor subtypes.

Interpreting High vs. Low Activity Scores

A high GSVA score indicates that many of the pathway’s genes are among the most highly expressed in that sample. This suggests the pathway is biologically active, or “turned on.”
A low GSVA score means the pathway genes are among the least expressed, suggesting that the pathway is inactive or “turned off.”
Scores are continuous, so you can see degrees of activation — not just “on/off” behavior.

Importantly, these scores do not depend on pre-defined sample groups (like treated vs. untreated). GSVA evaluates each sample individually, making it powerful for exploring heterogeneity, such as in tumor subtypes or patient stratification.

How It Relates to Up- or Downregulation

GSVA itself does not label anything as “up-regulated” or “down-regulated.” However, by comparing pathway scores between conditions or timepoints, you can infer regulatory shifts:

A pathway that consistently has higher scores in disease vs. healthy samples can be interpreted as up-regulated in disease.
Lower scores may indicate the pathway is down-regulated or inactive.

This approach is widely used in target discovery, biomarker research, and mechanism-of-action studies.

In Practical Terms

You might interpret pathway activity as a proxy for biological function:

High activity in a cell cycle pathway = cells are likely dividing.
High activity in an interferon signaling pathway = immune system is responding.
Low activity in a metabolic pathway = potential suppression or dormancy of energy processes.

In application to Pharma and Biotech research, this kind of pathway-level insight is key for:

Understanding mechanisms of action for drugs
Identifying which biological processes are driving disease
Tracking response to therapy at the systems level
Discovering biomarkers that reflect pathway engagement
Comparative biology across preclinical models and clinical samples

Of course, it's important to remember that GSVA scores are ultimately based in gene expression, so additional experiments are required to validate hypotheses of functional changes.

In summary, GSVA helps researchers move from normalized data to biologically meaningful insights, uncovering the processes that drive disease, drug response, or resistance.