Help Analysing Perturb-Seq Data - How do I extract 10X cell barcodes and relevant guides from fastq files?

Question

[jesswhitts]

Hello,

I am analysing some perturb seq data. For the single-cell, ECITE-seq was used, so we have GEX and protein expression values. The perturb-seq guide library was sequenced separately. I have managed to analyse the ECITE-seq part as normal, using cellranger and a CITE-seq workflow, but I'm not sure of the best way to integrate the perturb-seq guide information...

I think I need to somehow create a table containing the 10X cell barcode, 10X UMI, and alignment to a relevant sgRNA from the ones used. Any suggestions on how I can do this? I suppose I need to create a custom reference transcriptome for the sgRNA library and align my fastq files to it?

Best wishes, Jess

[lei]

excuse me, can you share how to use cellranger multi? Especially the config csv. Thanks!

[ATpoint]

Please read the CellRanger documentation, there is no point to repeat what they have written there unless you ask a specific question.

Answer 1

[yura.grabovska]

If you just want to extract relevant features like 10x barcode or UMI from fastq data you can process the fastq manually.

For example this will extract the 16-nt 10x barcode from a typical R1 by removing the last 10 characters from the 26-length read

cat Your_R1_001.fastq | \        #cat the R1 and pipe it to awk
awk '(NR%4==2)' | \              #take every 4th line of the fastq starting at line 2
sed 's/.\{10\}$//' > Your_R1.txt #empty the last 10 characters to only keep the 16-nt 10x barcode

If you know the structure of your R1/R2 then you can manipulate them as needed. But as you said, you probably should try a more sophisticated and robust approach

Answer 2

[jesswhitts]

Found a pretty simple solution in the end - using cellranger multi and rerunning my analysis with the CRISPR guides as features

[yura.grabovska]

Good to know! Thanks for updating with a solution