fix_batch_corrections

[byee4]

          Just taking some notes here

I found strong batch effect in polyA. I am thinking of the following plan:

1. salmon parameters: --seqBias --gcBias --posBias. The documentation also suggest checking GC bias from multiQC distributions https://salmon.readthedocs.io/en/latest/salmon.html
2. combat salmon quant.sf prior to running the next step of labrat.

Originally posted by @algaebrown in #4 (comment)

[byee4]

6c70ca3 contains --gcBias and --seqBias, but not --posBias. Should we add @algaebrown?

Also we are not running COMBAT as previously discussed