10X scRNA-seq: Demultiplex into Samples with I1, R1, and R2 FASTQs

Question

[QVINTVS_FABIVS_MAXIMVS]

I was given three files by a collaborator who is now on holiday and I'm looking for a quick answer for those who are not on holiday :)

I have three FASTQs from a 10x v2 scRNA-seq run.

The file with I1 contains what I assume is the sample index (8mer)

@E00527:118:HW5HWCCXY:7:1101:3315:1643 1:N:0:ACATTACT
ACATTACT
+
AA---<A-
@E00527:118:HW5HWCCXY:7:1101:3579:1643 1:N:0:ACATTACT
ACATTACT
+
AA<<--F<

The FASTQ with R1 seems to contain the cell barcodes and UMI

@E00527:118:HW5HWCCXY:7:1101:3315:1643 1:N:0:ACATTACT
GGACGTCCACATCCGGGCGGGTCGTCT
+
<AFF-AFFFJJ<FFJ-J<77FAAJ-A7
@E00527:118:HW5HWCCXY:7:1101:3579:1643 1:N:0:ACATTACT
GGTGAAGGTCATACGGTGTTTCTTTTT
+

The FASTQ with R2 seems to be the actual cDNA bit that was sequenced.

Am I correct with this so far?

So given the three files, how can I create a sample-specific FASTQ files. Sometimes the sequenced sample index has N nucleotides so it's not as simple as making a new FASTQ for each sample index.

I am expecting two samples in this FASTQ file... Thanks!

[GenoMax]

Not sure if what you have received is output of cellranger mkfastq or just plain bcl2fastq2? Sounds like the latter if there are more than one samples present. You can look at this page for further analysis options.

Eventually you will want to use cellranger count/aggr to do further analysis.

Answer 1

[swbarnes2]

Sometimes the sequenced sample index has N nucleotides so it's not as simple as making a new FASTQ for each sample index.

This is why you should not try to DIY this, but just use cellranger software. You'll need to find out from whoever made the library which kit was used, as you'll want to tell the software that. But there are only two options, so it's not onerous to just try them both.

Newer versions of cellranger don't care about the I1 file.

Cellranger also makes output that goes well into R's DropletUtils method of import, which is helpful.