Because GPSAdb contains 6,245 perturbation datasets with pre-computed differential-expression analyses and GSEA results, it can be used to back-trace the upstream regulators of a single gene or an entire gene set.

Querying Upstream Regulators of a Single Gene

Taking TP53 as an example, the Query module instantly returns all perturbations that significantly change TP53 expression.

For every “perturbed gene A → target gene TP53” pair, GPSAdb provides a ready-made expression plot that visualises the magnitude and direction of the change.

Querying Upstream Regulators of a Gene Set

If you want to discover which genes regulate a specific pathway or signature, eight built-in gene-set libraries are available.
This is a unique feature of GPSAdb: thanks to the large number of high-quality perturbation datasets, it can reliably identify upstream regulators for entire gene sets—an otherwise difficult task.

Custom Gene Sets

Need to analyse a user-defined gene set instead of our built-ins?
Simply switch to the BioTrigger page and upload your own list.

If you have a gene set of interest and want to identify which perturbations modulate the activity of that set, the BioTrigger module is designed for exactly this task.

Data Input Options

BioTrigger accepts two convenient input methods:

1. Copy & Paste
   Simply paste a newline-separated list of gene symbols into the text box.

   

2. File Upload
   Upload a one-column table (CSV/TSV) that contains your gene list.
   Make sure the file has a header—e.g. a column named geneSet.

   

Running BioTrigger

Click Run BioTrigger and wait ~20 seconds while the analysis is executed.
A new results interface will then appear.

• RankPlot displays the top-ranked genes whose perturbation most strongly affects your uploaded gene set.
• Select any row in the table to focus on a specific dataset, then click Plot to visualise:
  • an updated RankPlot limited to that gene (multiple datasets are shown if available), and
  • a GSEA plot illustrating how your gene set is enriched within the chosen dataset.

When you knock down a gene in a cell line—or treat the cells with a drug—you often want to know which other genes mediate the observed effect. Standard GSEA is not designed for this task.
fastGPSA fills this gap by finding perturbations whose transcriptional signatures resemble your own experiment.

How fastGPSA Works ⚙️

1. Input Upload the differential-expression results of your experiment (typically produced by DESeq2 or edgeR).
2. Gene-set construction fastGPSA automatically takes the top 150 up-regulated genes and the top 150 down-regulated genes and treats them as two custom gene sets.
3. Signature matching These two gene sets are then queried—via GSEA—against all 6,245 perturbation profiles in GPSAdb.
4. Ranking Perturbations that yield similar enrichment patterns are ranked at the top, revealing genes whose knock-out/knock-down produce effects most similar to yours.

Data Input

If your data can be used for GSEA, it can be used for fastGPSA—the required input is the same: an ordered geneList.

1. Copy & Paste

Paste two columns of data:
gene and ranking metric (usually log₂ FC).

2. File Upload

Upload a two-column table in CSV/TSV format.

Tip The table downloaded from the Browse page can be used here without any modification.

Running fastGPSA

Click Run fastGPSA. A notification will appear; the analysis takes 30–60 seconds.
You can freely explore other GPSAdb functions while you wait.

Result Interface

When the run finishes, a new dashboard opens:

• Left panel Top-ranked genes whose perturbation mimics your signature.
• Right panel Full list of genes that pass the similarity threshold.

Select any dataset you are interested in and click the green Details button.

Visualisations

• Rank Plot Shows where the selected gene(s) sit in the similarity ranking—multiple datasets for the same gene are displayed together.

• GPSA Plot Overlays the enrichment curves for your up and down gene sets within the selected dataset (e.g. LATS1 knock-out in MCF-7).

  • Green = genes up-regulated in both your data and the reference dataset.
  • Purple = genes down-regulated in both.

  This indicates a high resemblance between your perturbation and the reference (LATS1 KO), suggesting a potential regulatory relationship between LATS1 and ESR1.

• FuncPlot Compares GSEA results for the 50 Hallmark pathways between your dataset and the selected GPSAdb dataset.

  • Red dots = pathways enriched in both.
  • Grey dots = the remaining Hallmark pathways.

fastGPSA therefore gives you a brand-new way to dissect your data and uncover biologically meaningful connections. Whenever you can run GSEA, you should also try fastGPSA!

In this example, we demonstrate how to use the BROWSE module to analyse the RNA-seq data obtained after knocking down ESR1 in MCF-7 cells and how to extract additional biological insight from the results.

Exploring the dataset with Browse

A search for ESR1 returns seven datasets containing ESR1 perturbations.
Here we select the first entry, D21455.

Click Explore.
The principal-component analysis shows that PC1 explains 98.69 % of the variance, indicating a clear separation between control and siRNA groups.

In the differential-expression table, ESR1 is strongly down-regulated, confirming the siRNA knock-down efficacy.

Querying the downstream target GREB1 reveals a significant decrease in its expression, further supporting functional repression of ESR1.

Interim conclusions

• ESR1 expression is efficiently silenced by siRNA.
• The ESR1 target gene GREB1 is concomitantly down-regulated, indicating loss of ESR1 activity.
• A dominant PC1 suggests a pronounced cellular response to ESR1 knock-down.

Overall, the experiment appears highly successful. The next question is: through which pathways does ESR1 act in MCF-7 cells?

Inspecting GSEA results

Navigate to the GSEA tab.

The Hallmark analysis shows that two oestrogen-related pathways are suppressed, consistent with ESR1 inhibition.
In addition, HALLMARK_G2M_CHECKPOINT and HALLMARK_E2F_TARGETS—both linked to cell-cycle progression—are down-regulated, matching the observed reduction in MCF-7 cell proliferation after ESR1 loss.

The Browse module thus allows users to effortlessly explore any of the 6 245 perturbation datasets in GPSAdb and to derive meaningful biological insights with only a few clicks.

In this comprehensive example, we show how to use fastGPSA to analyse the RNA-seq data generated after knocking down ESR1 in MCF-7 cells and how to extract new biological insights from the results.

Using fastGPSA to pinpoint key regulators

First, select the ESR1 knock-down dataset in MCF-7 cells.

Next, export the geneList for D21455 and use it as the input for fastGPSA.

Submit the list to fastGPSA.

The fastGPSA output reveals several ESR1-related datasets at the top of the ranking.
Notably, LATS1, a YAP-pathway member known to maintain ESR1 function, is highlighted—confirming that fastGPSA retrieves both positive controls and findings reported in the literature.

Among the hits, OVOL2 is of particular interest: it has been implicated in breast-cancer progression but its connection to ESR1 is unclear.
The GPSA plot shows that genes up-regulated after ESR1 knock-down are likewise up-regulated after OVOL2 knock-down, whereas down-regulated genes behave similarly in both perturbations—suggesting comparable cellular effects.

Validating fastGPSA results with Browse

Because OVOL2 appears in the GPSAdb hit list, we can explore its dataset directly with the Browse module.

Browse confirms that OVOL2 knock-down leads to a reduction in ESR1 expression.

GSEA further shows that oestrogen-response pathways are significantly down-regulated after OVOL2 loss, indicating that OVOL2 modulates ESR1 activity and warrants experimental validation.