Patient bam processing

[pj-sullivan]

Closes #3

Control bams will be processed separately; this is just for the patient samples.
Once sbfs is configured following the README instructions, run:

cd analyses
bash 01-cram-to-bam.sh -i input/test-input.tsv -m input/manifest.tsv
md5sum -c results/bams/md5sum.txt

[rjcorb]

as discussed in Slack-- the shell script should be updated to take input/manifest arguments via command line to make this more robust. The input files provided here could be included as test files in repo.

[pj-sullivan]

Made the input files arguments, and added (optional) inputs for defining the columns with the information needed for the script, so input files with a different format can be used.

Also added an md5sum for the resulting files to confirm that the script is working as expected.

[rjcorb]

This ran successfully for me and the md5sums match yours. I made a few suggestions that I think will ensure this will be more robust against different input data.

[rjcorb]

I think we should be doing these queries by BS_ID rather than patient ID, because there are often multiple RNA-seq cram files from different sample collections from the same patient.

[rjcorb]

A few notes here:

• To make this more robust, I think this should be updated to consider both the start and end positions and make the windows around those. I think for variant cases only one position is relevant, but for splice events where the exons might be far apart, we'd want to make sure we're capturing a window around this interval for plotting.
• I think we might want to create a standardized file format for these input files to make sure these columns indices are always correct. I think the input file should include: ID, chr, start_pos, end_pos, gene_name (anything else?). And then you can build in a check to make sure any input file that is supplied has these columns in this order.