Signatures

Core Feature

This feature is available in all MATISSE Explorer deployments.

Signatures allow you to group multiple continuous variables together and analyze their combined values. While commonly used for genes, you can include any continuous feature column in your dataset. This is useful for visualizing pathway activity, cell signatures, or any custom variable combinations.

What is a Signature?

A Signature is a named collection of continuous variables that you define. You can include gene expression values, quality metrics, or any other continuous features. Instead of viewing one variable at a time, you can:

• Combine expression - Calculate aggregate scores across multiple genes
• Visualize patterns - See where gene sets are co-expressed
• Subset data - Find cells expressing your genes of interest
• Compare groups - Use signatures as axes or color in plots

Use Cases

• Immune signatures - Group immune markers (CD3E, CD4, CD8A, etc.)
• Pathway analysis - Collect genes in a biological pathway
• Cell type markers - Define marker genes for cell populations
• Custom signatures - Any combination relevant to your research

Creating a Signature

Step 1: Open Signatures Tab

1. In the Explorer sidebar, click the Signatures tab
2. You'll see existing signatures (if any) and a create option

Step 2: Click Create

1. Click the Create Signature button
2. A creation dialog opens

Step 3: Enter Details

Fill in the form:

Field	Description	Required
Name	Display name for the signature	Yes
Description	Optional notes about the signature	No
Variables	Variable names from continuous features (one per line or comma-separated)	Yes

Step 4: Add Variables

You can add variables in three ways:

Option A: Browse MSigDB gene sets

1. Click the Browse MSigDB gene sets... button
2. The MSigDB search dialog opens

3. Search for gene sets by keyword (e.g., "apoptosis", "BRCA1") - minimum 2 characters
4. Filter results by:

Filter	Options
Species	All, Human, Mouse (auto-detected from project data)
Category	All, Hallmark, Reactome, GO

5. Each result shows the gene set name, gene count, and species badge
6. Click Add genes on a gene set to add all its genes to your Signature
7. The Signature name and description are auto-filled if empty

TIP

Species is auto-detected from your project's variable naming conventions: all-uppercase names (e.g., BRCA1) indicate Human, while title-case names (e.g., Brca1) indicate Mouse.

Option B: Select from project

1. Click the Select variables from project... button
2. Browse or search available continuous features
3. Check the variables you want to include
4. Confirm your selection

This method ensures you only select variables that exist in your dataset.

Option C: Enter manually

Type variable names directly in the text area:

• One per line:

  CD3E
  CD4
  CD8A
  FOXP3

• Or comma-separated:

  CD3E, CD4, CD8A, FOXP3

This method is useful when pasting gene lists from external sources.

Step 5: Save

Click Create or Save to create the signature. It appears in your list immediately.

Creating from DEG Results

InSilicoLab Feature

This method is available when using DEG analysis in InSilicoLab.

You can create Signatures directly from DEG analysis results:

1. Run a DEG analysis and open the results
2. In the Data Table, select genes of interest using checkboxes
3. Click Create Signature in the table toolbar
4. Enter a name and optional description
5. The selected genes are added as variables

See DEG Analysis - Create Signature from Results for details.

Variable Matching

When you create a signature, MATISSE Explorer checks which variables exist in your dataset.

Matched Variables

Variables that exist in your data and can be used for calculations.

Unmatched Variables

Variables not found in your data. These are:

• Ignored in calculations
• Listed for your reference
• May indicate typos or missing data

When creating a Signature with unmatched variables, a confirmation dialog appears listing the unmatched variables and asking whether to proceed. You can continue with only the matched variables or cancel to fix the variable names.

Tip: Check spelling and naming conventions if variables don't match.

Combination Methods

When using a Signature for visualization or subsetting, choose how to combine the variables:

Method	Formula	Best For
Max	Maximum value among genes	Cells expressing ANY gene highly
Min	Minimum value among genes	Cells expressing ALL genes
Sum	Sum of all gene values	Total pathway activity
Average	Mean of gene values	Balanced representation
UMI Count	ln(10000 × sum/totalUMI + 1)	Normalized expression

Choosing a Method

• Max: Good for marker discovery - finds cells where at least one gene is high
• Min: Good for co-expression - finds cells where all genes are expressed
• Sum/Average: Good for overall pathway activity
• UMI Count: Good for comparing across cells with different sequencing depth

Using Signatures

For Color Mapping

1. Select a Signature from the list
2. Choose a combination method
3. The map colors by the computed score
4. Adjust range and colormap as needed

For Subsetting

Add a Signature condition in the Subset:

1. Open the Subset Editor
2. Add a new condition
3. Select Signature as the condition type
4. Choose your signature and combination method
5. Set comparison (e.g., score > 1.0)

For Scatter Plot Axes

Use Signatures as X or Y axis in scatter plots:

1. Open the Plot extension
2. In axis configuration, select a Signature
3. Choose the combination method
4. The plot uses the computed scores

Managing Signatures

Edit a Signature

1. Click on the Signature in the list
2. Click the Edit button (pencil icon)
3. Modify name, description, or variables
4. Save changes

Delete a Signature

1. Click on the Signature in the list
2. Click the Delete button (trash icon)
3. Confirm deletion

Warning: Deletion is permanent. Any subsets using this signature will need updating.

View Details

Click on a Signature to see:

• Full list of variables
• Match status for each variable
• Quick actions (edit, delete)

TIP

The Signatures tab displays the variable count next to each Signature name, making it easy to see how many variables are in each set at a glance. When many Signatures exist, the tab scrolls horizontally rather than wrapping.

Tips for Working with Signatures

Organizing Signatures

• Use descriptive names (e.g., "T cell markers" not "Signature 1")
• Add descriptions for future reference
• Group related genes logically

Optimizing Performance

• Large signatures (50+ genes) may take longer to calculate
• Results are cached for faster subsequent access
• Pre-compute important signatures before intensive analysis

Validation

• Review unmatched variables for typos
• Check gene naming conventions in your dataset
• Use official gene symbols when possible

Examples

Immune Cell Signature

Name: T Cell Markers Variables:

CD3D
CD3E
CD4
CD8A
CD8B

Method: Max (for any T cell marker positive)

Proliferation Score

Name: Proliferation Genes Variables:

MKI67
TOP2A
PCNA
MCM6

Method: Average (balanced proliferation score)

Custom Pathway

Name: Interferon Response Variables:

ISG15
MX1
OAS1
IFIT1
IFITM1

Method: Sum (total pathway activity)

Troubleshooting

No Variables Match

• Verify gene names are correct
• Check if your dataset uses different naming conventions
• Try searching in Continuous features for available names

Scores Look Wrong

• Check combination method is appropriate
• Verify matched variables are the ones expected
• Look for data quality issues in source features

Performance Issues

• Reduce variable count if possible
• Wait for caching on first computation
• Consider subsetting data first to reduce calculations

Related Topics

• Color Mapping - Use signatures for coloring
• Subset - Subset by signature scores
• Scatter Plot - Use signatures as axes