Analyse ASEReadCounter output

[LGGray]

Hi there, thank you for putting out this tool!

I was wondering if it could be possible to run brewerIX on the output of ASEReadCounter? As I am working on many scRNA-seq samples it is more efficient to run the GATK pipeline on a HPC.
If this feature is difficult to implement, could you please point me towards the specific R scripts used to handle the ASEReadCounter output files?

Thank you in advance,

Lachlan

[cavei]

Hi,
Thank you for your interest in BrewerIX.
Unfortunately there is no such possibility.

To be used in HPC I suggest creating a Singularity Image File.
I’ll provide a Docker file to create a docker image that can be converted in singularity.

That’s the best choice.

Alternatively, what you can do is to manually, i.e. create the ASER_table.txt file for your dataset and hijack BrewerIX.

Briefly, starting from the sample specific ASE table,
you can extract and collapse with “_” the columns:
contig position variantID refAllele altAllele
to obtain something like 10_119726283_rs1554881876_G_A and then collapse refCount and altCount and the binomial test in the form refCount,altCount,binTest to obtain somthing like 5,0,1.0 (keep all 1.0 values if you don’t need the pvalues).

Then you need to merge all the samples… indicates missing with NA.

You need to end up with an ASER_table.txt file that look like this:
chr_pos_rs_ref_alt HPD01_naive BJ_fibroblasts HPD00_primed
10_119726283_rs1554881876_G_A 5,0,1.0 NA NA
10_119726284_rs984606804_G_A 5,0,1.0 NA NA
10_119726286_rs768504947_C_T 5,0,1.0 NA NA
10_119726287_rs1326272336_C_A 5,0,1.0 NA NA
10_119726289_rs1337074486_C_T 6,0,1.0 NA NA

where HPD01_naive BJ_fibroblasts and HPD00_primed will be you sample names (you can add as many columns as your samples.

Then to hijack BrewerIX you need to create a dir with a fake fq file (or paired; with at least one of the names presented in the columns of the ASER_table.txt) and respective fake bam (including index).

For example:
mkdir example01
cd example01
touch HPD01_naive_1.fq.gz HPD01_naive_2.fq.gz HPD01_naive.bam HPD01_naive.bam.bai

Create an empty file name do_not_run_ASER (touch do_not_run_ASER) and finally copy in the dir example01 the ASER_table.txt file you have create with yous samples.

Then you can run BrewerIX from example01 directory and brewerIX will give you the brewer_table and the graphical output.

Hope this will do the trick.
If you have errors or questions, do not hesitate to contact me.

[LGGray]

Hi Cavei,

Thank you for your detailed instructions!
I have made ASER_table.txt with the combined cells of each individual as columns and included 1.0 for all SNPs. However, as the .bam and bam.bai files are empty, BrewerIX was unable to index the files and failed. To circumvent this, I have used a random .bam file from one of the cells and it works!

Though I understand that github isn't the typical place for discussion, I am curious how to determine reactivation of the X chromosome from the brewer-table? My main interest is escape from XCI and skewed XCI.

Thank you again for your help!

Lachlan

[cavei]

Hi Lachlan,
glad it works.
For XCI escapee and skewed XCI since you are in single cell you "just" need to quantify the bi-allelic SNP from the X.
Since looking through the BrewerIX interface may be difficult due to the large size of the image, I have written an R package with functions that eat the brewer-table and can summarize output for single cell.
You can find the package here https://github.com/Romualdi-Lab/rbrewerix

In attachment a small reproducible example with some explanation that I hope will help you through.
sc-test-data-brewerIX.zip

Otherwise I will be glad to help.

Cheers