Add cloud compatibility

[bkohrn]

Long-term goal, placed here to enable tracking. An eventual goal would be to make this pipeline into something that can be easily run on the cloud. This will require several steps, and may include (among other things):

• A transition to docker containers instead of using conda environments
• Changing the main pipeline engine from Snakemake to something else that is easier to set up and run on the cloud. While I know that Snakemake can be run on the cloud, I don't know how easy it is, and my impression is that it will require a significantly different setup on cloud vs. local systems.
• Other changes to aid upload of raw data to the cloud and retrieval of processed data from the cloud, including potential reruns. This may include compressing the run directory after completion.