Add `batch` identifier to cohorts & support estimation of batch effects

[jburos]

Would be useful to keep track of batch identifiers so we can estimate batch effects as part of standard protocol.

[jburos]

Two parts to this request:

1. [ ] add batch to sample meta-data
2. [ ] implement method for summarizing batch effects, depending on type of sample

Possible methods for detecting batch effects:

1. run combat on our samples to estimate batch effects from RNA-seq
2. use SVA to try to estimate batch effects
3. look for hierarchical clustering in DNA/RNA more generally