Multipe tables as input - sample_table

[KuechlerO]

Hi guys,

thx for your great tool!

I'm using PEP in my snakemake pipeline since some time (see here for docs: https://snakemake.readthedocs.io/en/stable/snakefiles/configuration.html#configuring-scientific-experiments-via-peps).
Yet, I have the problem, that my pipeline takes a consistent list of control samples, while my list of samples to compare against varies all the time.
So I was wondering, how I can use PEP to combine in each project both:

1. Control samples (e.g. control_samples.csv)
2. Samples to analyse (e.g. analysis_samples.csv)

As far as I understand, with sub-projects (https://pep.databio.org/spec/howto-amendments/) I can only define different variable settings. But I cannot combine many sample sheets into one.
Or have I missed a point?

[nsheff]

Thanks for raising this question.

Right now, PEP doesn't have a built-in way to concatenate multiple sample sheets into one combined sample table automatically. Here are some ideas:

1. Can you concatenate the 2 CSV files (use a column to indicate sample_type as control or analysis)? I guess the problem with this is that you'd need to duplicate the control samples? Could you do that in the pipeline itself?
2. Can you make the pipeline take 2 PEPs, and then just always pass the control ones, as a separate file? In other words, haev the pepfile point to analysis_samples -- this is what varies. Then, just read in the control sample pep in your pipeline using python directly (not as a snakemake pepfile). eg:

import peppy
control_pep = peppy.Project("control_samples.csv")

[nsheff]

Just to add -- you're right that sub-projects is not the right answer here.