basic descriptions of what's being done and why

[roblanf]

We are making progress! But one thing I've noticed is that we spend a decent amount of time trying to decipher what should be happening if everything works.

So, can I ask you to add two sections to the readme at the end of each of the scripts, titled something like:

What it does

and

Expected output

E.g. for the clean_reads.sh in the first section we'd want stuff about RCorrector, khmer, and the rationale for combining them here. In the section section we'd want to know what files we'd expect to see in the output directories (as well as subdirectories that should be empty, etc), and a sentence on what each file should contain.

Thanks!

R

[adamjorr]

I will work on this. Thanks for the feedback, it helps to know what is confusing so I can make the documentation as helpful as possible.

Specifically for clean_reads, the khmer and rcorrector steps are combined merely for ease of use. Do you think they should be separated? Maybe I could spin them into two separate scripts and keep clean_reads as a wrapper for them both. Since we're considering ditching khmer anyway, that may be easier for maintainability. It may also be easier to turn the pipeline into a make file, which is something @reedacartwright wants me to do.

[roblanf]

The issue for me at least is that without knowing exactly why you're using rcorrector and khmer in the first place, it's hard to know what I think...

Starting with the descriptions of what and why would be a good start, and then we could worry about maintainability later.

On 21 February 2017 at 04:16, Adam Orr notifications@github.com wrote:

I will work on this. Thanks for the feedback, it helps to know what is confusing so I can make the documentation as helpful as possible.

Specifically for clean_reads, the khmer and rcorrector steps are combined merely for ease of use. Do you think they should be separated? Maybe I could spin them into two separate scripts and keep clean_reads as a wrapper for them both. Since we're considering ditching khmer anyway, that may be easier for maintainability. It may also be easier to turn the pipeline into a make file, which is something @reedacartwright https://github.com/reedacartwright wants me to do.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/adamjorr/somatic-variation/issues/23#issuecomment-281135537, or mute the thread https://github.com/notifications/unsubscribe-auth/AA2pE7lpaO-uMyOLbCWG47La7YtzPSv-ks5recpfgaJpZM4MFwdR .

-- Rob Lanfear Ecology, Evolution, and Genetics, The Australian National University, Canberra

www.robertlanfear.com

[adamjorr]

I updated the docs, let me know if it helps