Skip to content

Cell Analysis Guide

This comprehensive guide covers all aspects of automated cell analysis in napari-mAIcrobe, from basic morphometry to cell classification.

🎯 Overview

The "Compute cells" widget provides:

  • Morphological analysis β€” Shape and size measurements
  • Intensity analysis β€” Fluorescence quantification
  • Cell classification β€” Deep learning classification with default models for cell cycle phase determination in S. aureus
  • Colocalization analysis β€” Multi-channel correlation
  • Report generation β€” HTML output

πŸ§ͺ Analysis Workflow

Step 1: Prepare Your Data - Segmented cells (Labels layer) - Image channels: membrane and DNA as needed

Step 2: Configure Analysis Parameters

Essential Settings - Label Image: segmentation (required) - Membrane Image: fluorescence channel - DNA Image: DNA staining (optional) - Pixel size: ΞΌm/pixel (optional)

Subcellular Segmentation - Inner mask thickness (default: 4) - Find septum / Find open septum - Septum algorithm: Isodata or Box

Fluorescence Analysis - Baseline margin (default: 30)

Cell Cycle Classification - Enable classification; select model or custom path - Custom model input: Membrane, DNA, or Membrane+DNA - Custom model max size (default: 50 px)

Options - Compute Colocalization - Generate Report (+ Report path) - Compute Heatmap


πŸ“ Morphological Measurements

napari-mAIcrobe computes shape and size parameters using scikit-image regionprops.

Basic Shape Parameters - Area (px, ΞΌmΒ² with pixel size) - Perimeter - Eccentricity


πŸ’‘ Intensity Analysis

Quantify fluorescence signals in subcellular compartments.

Channel Measurements - Baseline intensity (subtracted from other measures) - Cell Median - Membrane Median - Cytoplasm Median - Septum Median (if detected) - Fluorescence Ratios (100%, 75%, 25%, 10%) when septum enabled - DNA Ratio if DNA channel provided


🧠 Cell Classification

Use deep learning models to automatically classify cells.

Pre-trained Models - 6 specialized models for S. aureus: DNA+Membrane (Epi/SIM), DNA (Epi/SIM), Membrane (Epi/SIM)


πŸ”— Colocalization Analysis

Quantify spatial relationships between two fluorescence channels using Pearson correlation for Whole Cell, Membrane, Cytoplasm, and Septum (if available).


πŸ” Interactive Filtering

Use the "Filter cells" widget for real-time quality control: 1. Select Labels layer 2. Add filters for any measured feature 3. Preview filtered population 4. New "Filtered cells" layer contains only the selected cells


πŸ“„ Reports

Enable Generate Report to create an HTML report with per-cell images and a CSV with all properties.

Further reading: Cell Classification Guide, API Reference, Tutorials.