distiller

[yardenmatok203]

Hey,

can you please elaborate your preprocessing with distiller and cooler?

Thanks!

[gfudenberg]

Distiller code for going from reads to binned contact maps can be found here:
https://github.com/open2c/distiller-nf
(it relies on pairtools: https://www.biorxiv.org/content/10.1101/2023.02.13.528389v1.full and cooler)
We used defaults for filters, mapq30 filter for read quality, and set the binsize used for cooler to 2048bp. This was chosen as a multiple of 2 close to the maximum resolution we felt comfortable analyzing the contact datasets in the paper given their read coverage and experimental protocol details.

Use of cooler is integrated into the distiller pipeline. If you already have mapped reads in a .pairs format, though, you can use cooler cload: https://cooler.readthedocs.io/en/latest/cli.html#cooler-cload.

[yardenmatok203]

thank you!
Did you used raw files from: https://data.4dnucleome.org/files-fastq/4DNFITHCIIBX/ ?

[gfudenberg]

Yes, but for that experiment I'd just start with the .pairs file, as they were processed with distiller anyways.

[yardenmatok203]

This will do the ICA and binning?
please can you tell me which parameters should I use in order it will be close to your processing?

Thank you,
Yarden

[gfudenberg]

Hi Yarden,
cooler handles the iterative correction (ICE) and binning, we used defaults for human data-- please see the cooler docs/manuscript for more info: https://cooler.readthedocs.io/en/latest/datamodel.html, https://academic.oup.com/bioinformatics/article-abstract/36/1/311/5530598
Hope that helps!

[yardenmatok203]

I want to use .hic data, what should I use?

thanks a lot!!!