Gene Expression Visualization

The visualization module allows you to explore gene expression patterns across your single-cell dataset by plotting specific genes of interest in reduced dimensional space (UMAP or PCA).

Step-by-Step Visualization Process

• (1) Select Dimension Reduction: Choose between UMAP or PCA for your visualization space
• (2) Search Genes: Type the genes of interest in the search field
• (3) Review Selection: The selected genes will appear on the right side for confirmation
• (4) Run Analysis: Click on Run to start the visualization process
• (5) Continue: Click on Continue to Results to proceed to the download section

Understanding Visualization Results

After running the visualization, you will see gene expression plots overlaid on your chosen dimensional reduction space. Each plot shows:

• Cell positions: Based on UMAP or PCA coordinates
• Expression levels: Represented by color intensity
• Spatial patterns: Revealing which cell populations express your genes of interest

This spatial representation helps identify cell types, validate clustering results, and understand the biological context of your single-cell data.

Gene Observations Plots

Use the Gene Observations container to color the embedding by per-gene metadata stored in var (e.g., total counts per gene, HVG flags). Configure:

• (1) Dimension Reduction: Choose UMAP or PCA as the display space.
• (2) Search: Filter available var columns by name.
• (3) Select columns: Pick one or more var fields to visualize.
• (4) Run: Generate one plot per selected column.

Each resulting plot colors cells according to the selected per-gene metric aggregated over the embedding. Use this to inspect HVG selection, gene-level filters, or feature-wide summaries.

Cell Observations Plots

Use the Cell Observations container to color the embedding by per-cell metadata stored in obs (e.g., total counts, percent mitochondrial, Leiden cluster, user labels). Configure:

• (1) Dimension Reduction: Choose UMAP or PCA as the display space.
• (2) Search column: Filter available obs columns by name.
• (3) Select columns: Pick one or more obs fields to visualize (numeric or categorical).
• (4) Run: Generate one plot per selected column.

Plots use continuous scales for numeric columns and discrete palettes for categorical columns (e.g., clusters/labels). This is ideal for QC review and for verifying labeling decisions before downstream analyses.