Project Nicolas IHC Checklist

[irinabelaia]

Samples:

• 24 FFPE tumor sections that were immunohistochemically stained for three proteins: HER2, HER3 and SORLA
• 48 images (24 for the HER2/SORLA staining, 24 for the HER2/HER3 staining)
• the cancer parts keep their epithelial features and present two polarity phenotypes (apical-in and apical-out)
• Mix 1:
  • HER2 = Cy3 / Ch2
  • SORLA = AF647 / Ch3
• Mix 2:
  • HER2 = Cy3 / Ch2
  • HER3 = AF647 / Ch3

Requested analysis:

• Measure the intensity of each signal (SORLA, HER2, HER3) in the whole cancer region and in the epithelial cancer region (apical pole and basal pole)

Tasks:

• [x] Improve ROIs/masks for apical-in by drawing manual ROIs
• [x] Recalculate the intensity analysis in all requested regions with the updated ROIs
• [x] Correlation analysis of intensity HER2 vs HER3 and HER2 vs SORLA within on patient samples for 1) whole region, (apical-in and apical-out), and for 2) basal and apical pole zones separately. Plot the correlations for each individual ROI and for the average per patient
• [x] To compare the intensity between zones, normalize the intensity value to ROI areas
• [x] In addition to the normalization by ROI area, perform the normalization by cell number within the ROI (segmentation and counting the cell nuclei within ROIs)? // Nicolas asked 29.04.2024
• [x] Investigate mean intensities over the entire data, comparing Zeiss- and BioFormats-derived images.

Inside note: In vitro there is a correlation between HER2 and HER3 (for apical-in) and the overall intensity lower in apical-out in comparison to apical-in

[junelsolis]

• [x] Rerun pipeline to include ROIs that were left out @junelsolis

Notes from meeting with Johanna 2024-04-19

• [ ] Look into making the plots interactive
• [ ] Separate the data in quartiles
• [x] Generate ROI images in folders for quick reference
• [x] Equalize XY axes in plots.
• [ ] Plot intensity distribution between ap-out and ap-in.
• [x] Generate data with intensity normalized to area