How to adjust for transcript length

[hammer]

@hammer commented on Sat Dec 17 2016

Another one from @jburos's 12/12 status report. Can we fill out this one a bit too?

[jburos]

At a high level, there are two ways to approach this:

1. Model number of reads as a function of transcript length; incorporate this adjustment into the model
2. Use a metric (alternative to # reads) that already adjusts for transcript length

In a way, this may not matter at all to our use case since we aren't comparing reads across different transcripts within a sample; instead we are always comparing the same transcript across samples and so the length may not matter. But it would be good to confirm this.