DeepTumour

The DeepTumour algorithm predicts the tissue of origin of a tumour based on the pattern of passenger mutations identified by Whole Genome Sequencing (WGS).

Running DeepTumour

Requirements:

• hg19 reference (.fa & .fa.fai)

Docker

1. Study the command options:

   docker run --rm ghcr.io/lincolnsteinlab/deeptumour:latest --help

2. Run DeepTumour (assuming VCFs and reference are under the current working directory):

   docker run --rm -a stdout -a stderr -v .:/WORKDIR ghcr.io/lincolnsteinlab/deeptumour:latest [OPTIONS] --stdout > [OUTPUT_JSON]

Singularity

1. Study the command options:

   singularity run docker://ghcr.io/lincolnsteinlab/deeptumour:lastest --help

2. Run DeepTumour (bind paths if needed):

   singularity run docker://ghcr.io/lincolnsteinlab/deeptumour:latest [OPTIONS]

Local Environment

Requirements:

• Python 3.9
• Pip

1. Install Python dependencies:

   pip install -r ./requirements/requirements.txt

2. Study the command options:

   python DeepTumour.py --help

3. Run DeepTumour:

   python DeepTumour.py [OPTIONS]

Appendix

About the DeepTumour Algorithm

Frequently Asked Questions

Publications

• Jiao, W., Atwal, G., Polak, P. et al. A deep learning system accurately classifies primary and metastatic cancers using passenger mutation patterns. Nat Commun 11, 728 (2020). https://doi.org/10.1038/s41467-019-13825-8
• Lincoln David Stein, Wei Jiao, Gurnit Atwal, Quaid Morris; Abstract 5046: DeepTumour: Identify tumor origin from whole genome sequences. Cancer Res 15 June 2022; 82 (12_Supplement): 5046. https://doi.org/10.1158/1538-7445.AM2022-5046