This is a repository for the cloneHD submission to the SMC-Het Challenge. It solves the following sub-challenges:
- Sub-challenge 1A - Predicting normal contamination
- Sub-challenge 1B - Predicting number of subclones
- Sub-challenge 1C - Predicting subclone proportions
- Sub-challenge 2A - Determining mutation assignments to subclones
- Sub-challenge 2B - Co-clustering matrix
The method is described in a published manuscript and on the project wiki. The results of the challenge will become available on th leaderboards in the coming months.
To test the workflow locally, run the shell script:
$ cloneHD_tool.sh --vcf Tumour1.mutect.vcf --cna Tumour1.battenberg.txt --sample Tumour1
You can modify the default number of trials providing the optional flag --trials <int>, and the number of independent optimizations with the optional flag --restarts <int>. You can also set the seed with --seed <int>.
You first have to follow these instructions to sign up on Google Compute Engine, set up your working environment and set up your VM instance.
If you have the Google Cloud SDK installed locally, you can SSH into the VM instance with gcloud:
$ gcloud compute ssh ubuntu@planemo
To clone this repository, run the following command in a local directory. For example:
$ cd /opt/galaxy/tools
$ git clone https://github.com/ivazquez/cloneHD-SMC-Het.git
To set up the workflow in a container using docker:
$ cd cloneHD-SMC-Het
$ planemo docker_build .
This will build and compile cloneHD plus all dependencies.