Cancers used

• (STAD) Stomach adenocarcinoma
• (COAD) Colon adenocarcinoma
• (LAML) Acute Myeloid Leukemia
• (BRCA) Breast invasive carcinoma
• (THCA) Thyroid carcinoma
• (BLCA) Bladder Urothelial Carcinoma

Disclaimer

Running the recognizer requires significant computational resources.

Usage

Install requirements

pip install -r requirements.txt

Data preparation

To prepare the data yourself, use this command to run the data preparation pipeline:
snakemake -s ./src/data_preparation/Snakefile --cores 1

To use already prepared data run the data_download script:
python src/data_download.py

It will download the data from a public repository to the appropriate folders.

Running the holdout classifier

Summed Embeddings

snakemake -s ./src/classifier_holdout/summed_embeddings_generation/Snakefile --cores 1

Models

snakemake -s ./src/classifier_holdout/models/Snakefile --cores 1

Plots

snakemake -s ./src/classifier_holdout/plots/Snakefile --cores 1

Latent Dimensions Classifier

Summed Embeddings

snakemake -s ./src/classifier_latent_dims/summed_embeddings_generation/Snakefile --cores 1

Models

snakemake -s ./src/classifier_latent_dims/models/Snakefile --cores 1

Plots

snakemake -s ./src/classifier_latent_dims/plots/Snakefile --cores 1

Subtype Classifier

Summed Embeddings

snakemake -s ./src/sub_type_classifier/summed_embeddings_generation/Snakefile --cores 1

Models

snakemake -s ./src/sub_type_classifier/models/Snakefile --cores 1

Plots

snakemake -s ./src/sub_type_classifier/plots/Snakefile --cores 1

TMB Classifier

Summed Embeddings

snakemake -s ./src/tmb_classifier/summed_embeddings_generation/Snakefile --cores 1

Models

snakemake -s ./src/tmb_classifier/models/Snakefile --cores 1

Plots

snakemake -s ./src/tmb_classifier/plots/Snakefile --cores 1

Recognizer

Summed Embeddings

snakemake -s ./src/recognizer/summed_embeddings_generation/Snakefile --cores 1

Models

snakemake -s ./src/recognizer/models/Snakefile --cores 1

Plots

snakemake -s ./src/recognizer/plots/Snakefile --cores 1