the google life sciences api creates a new snakemake container to run each of our jobs
so our dependencies must be installed anew by snakemake every time it runs a job, since each job is run in a new, empty container
this process may add substantially to the runtime of some of our instances and the cost of using them
we have two options:
1) separate our dependencies into different env files to install only the necessary dependencies for each job
2) create a custom snakemake container with our dependencies already installed in it and force snakemake to use our container instead of its own
the google life sciences api creates a new snakemake container to run each of our jobs so our dependencies must be installed anew by snakemake every time it runs a job, since each job is run in a new, empty container this process may add substantially to the runtime of some of our instances and the cost of using them
we have two options: 1) separate our dependencies into different env files to install only the necessary dependencies for each job 2) create a custom snakemake container with our dependencies already installed in it and force snakemake to use our container instead of its own