Implement bam-readcount CRAM support

[jasonwalker80]

Add support to all relevant CWL pipelines and subworkflows to use CRAM as an input format. This uses a new version of bam-readcount that natively supports CRAM.

[malachig]

I see no evidence of any work being done on bam-readcount itself:
https://github.com/genome/bam-readcount

If this is being tracked elsewhere, why?

[tmooney]

There is a branch on a fork where work has been taking place: https://github.com/seqfu/bam-readcount/tree/samtools-1.10, but it hasn't been pushed back to the main repository as of yet.

If I were to speculate as to why it's not tracked on the repo, it's related to the fact that this has been a hackathon-related topic and so worked on in bursts primarily using the more realtime communications of Slack (in the #bam-readcount channel). A lot of our issue-tracking has been coordinated around the hackathons as seen in this issue, so even things that reach beyond the analysis-workflows have mostly been coordinated via boards like here: https://github.com/genome/analysis-workflows/projects

[chrisamiller]

bam readcount 1.0.0 has full cram support. All that is needed is to upgrade the docker container in the relevant steps at this point