task1

Build your model

1. Go to the one of example models folder:

   • r-simple: [R] model by multiplying random numbers
   • r-magic: [R] model using Rmagic package from MAGIC
   • py-deepimpute: [Python] model using deepimpute

2. Build the Docker image docker.synapse.org/<synapse-project-id>/<model-name>:<image-tag>, like the following command:

   # in an example model folder
   docker build -t docker.synapse.org/syn123/example-model-task1:v1 .

Test the model locally

1. Find your training data. You could use seurat_pbmc3k_counts.csv, which is the count matrix from Seurat PBMC 3k dataset.

2. Create an output folder

   mkdir output
   

3. Run the dockerized model.

   docker run \
       -v $PWD/<training-data-folder>/:/input:ro \
       -v $PWD/output:/output:rw \
       docker.synapse.org/syn123/example-model-task1:v1

4. The imputed files (*_imputed.csv) should be saved to output/.

   $ ls output/*_imputed.csv
   

Submit this model to the DREAM Challenge

If your model meets the requirements and it can be pushed up to your project on Synapse. Please see the instructions on how to submit the model to synapse. Then submit it to the Task 1 of the DREAM Challenge.