Include a limitations of the study section

[jaclyn-taroni]

Off the top of my head, I would expect this to contain statements about:

• RNA-seq library strategy (related: Add text about RNA-seq batch effects to discussion section as part of limitations of the study #422)
  • Edit: also sequencing center (related: Revise S7 to include information about RNA-seq batch effects OpenPBTA-analysis#1691 and Differences in mitochondrially-encoded gene expression between RNA-seq samples sequenced at BGI@CHOP vs. NantOmics OpenPBTA-analysis#1601)
• Unbalanced sample size with respect to cancer_type seems worth mentioning more generally.
• Low tumor purity samples (Add information about thresholded tumor purity analysis to the manuscript #427)
• In the dataset in this paper, there were not enough longitudinal samples such that we felt comfortable performing analyses (https://alexslemonade.github.io/OpenPBTA-analysis/analyses/sample-distribution-analysis/03-tumor-descriptor-and-assay-count.nb.html#paired_genomic,_transcriptomic_assays). It is worth mentioning that more longitudinal samples will be available in the future (and how/where!).
• We might also reiterate why we took the independent specimens approach. (To me, this seems naturally to tie into the point above this one.)