This is a project for my paper

---

TODO list

---

Steps

processSvs.py => main.py

processSvs.py

Accepts meta data (Download meta) of TCGA Svs files. The svs files and the meta data should place in the same directory. This script will crop all svs to patches and filter patches by color, based on Lab color space. The patches are stored in a specific directory in the form <TCGAID>_pos.jpg.

main.py

Process all patches by Cellprofiler, count cell number of all patches. Select top 20 (default, can be set) patches with the highest cell density. Results are stored in a specific directory.

Records

run pipeline command:

python run_pipeline.py data/20x/hand-select-nosplit/ data/pipeline tangbo.cpproj -b 100

The cellprofiler project tangbo.cpproj has more features. Using a measurement generate by our script. Which could be easily change to the default method from cellprofiler.

Paper

The paper has been published in JBHI titled "A Novel MKL Method for GBM Prognosis Prediction by Integrating Histopathological Image and Multi-omics Data"

cite:

Zhang, Ya, et al. "A Novel MKL Method for GBM Prognosis Prediction by Integrating Histopathological Image and Multi-omics Data." IEEE journal of biomedical and health informatics (2019).