Nuclei segmentation in low magnification images of tumor tissue

The aim of this project is to develop a pipeline to automatically segment millions of nuclei in tumor tissue scans acquired at low magnification (20x). The pipeline is divided in: (1) A quality control step to (2) Segmentation step using Ilastik, an open source pixel classifier software (3) A custom script to perform watershed segmentation. We applied the pipeline to DAPI stained tissue sections of a cohort including lung adenocarcinoma and lung squamous cell carcinoma patients.

Online datasets

  1. experiment 10, sample 52
  2. experiment 13, sample 42

These data have been generated as part of an ongoing effort in our lab to build a large Cancer Nucleus Atlas and explore the spatial arrangement of chromatin and various sub-nuclear organelles and compartments in different tumor types. The Cancer Nucleus Atlas is funded by the Swedish Foundation for Strategic Research, the Swedish Cancer Society and the Cancer Research KI Program