This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Only two fastqs for 10X scATAC-seq data deposited on GEO database.

I am trying to replicate the upstream process of ATAC sequencing in an article (SRR26674024), but the uploaded raw data only contains two fastq files like SRR26674024_1.fastq.gz and SRR26674024_2.fastq.gz. I'm just wondering how to let the two files adapted to cellranger pipline. The first sequencing fragment of the two files shown as follows: @SRR26674024.1 A00869:659:H2M35DRX2:1:2101:29170:1000/3 NTACAAACTAAATGTGCCCTTGATGGAAAAGGAACCTTATGGAATGATGA + ?????????????????????????????????????????????????? and @SRR26674024.1 A00869:659:H2M35DRX2:1:2101:29170:1000/4 GCTGAGGCGGGAGAATCAGTTGAACCCAGGGGGCAGAGGTTGCAGTGAGC + ??????????????????????????????????????????????????

scatac-seq cellranger single-cell 10x

1 answer

Unfortunately 10x data on SRA is bit of a hit and miss thing.

It is possible to get the four files using fastq-dump by doing

$ fastq-dump -F --split-files SRR26674024

This should give you

$ more SRR26674024*fastq
::::::::::::::
SRR26674024_1.fastq
::::::::::::::
@A00869:659:H2M35DRX2:1:2101:29170:1000
AAACTCAT
+A00869:659:H2M35DRX2:1:2101:29170:1000
FFFFFFFF
::::::::::::::
SRR26674024_2.fastq
::::::::::::::
@A00869:659:H2M35DRX2:1:2101:29170:1000
GGGGTGCCCGCGGTAA
+A00869:659:H2M35DRX2:1:2101:29170:1000
,FFFFFF,FFFFF:FF
::::::::::::::
SRR26674024_3.fastq
::::::::::::::
@A00869:659:H2M35DRX2:1:2101:29170:1000
NTACAAACTAAATGTGCCCTTGATGGAAAAGGAACCTTATGGAATGATGA
+A00869:659:H2M35DRX2:1:2101:29170:1000
#FFF:FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF:FFF,FFFFFF:F
::::::::::::::
SRR26674024_4.fastq
::::::::::::::
@A00869:659:H2M35DRX2:1:2101:29170:1000
GCTGAGGCGGGAGAATCAGTTGAACCCAGGGGGCAGAGGTTGCAGTGAGC
+A00869:659:H2M35DRX2:1:2101:29170:1000
FFFFFFFFFFF,FFFFFFFFFFF:FFFFFFFFFF,FFFFFFFFFFFFFFF

These files are available for download via cloud SRA (e.g. https://trace.ncbi.nlm.nih.gov/Traces/?view=run_browser&acc=SRR26674024&display=data-access under original data) but unfortunately that process would not be free since you will need access to a AWS/GCP cloud bucket and a paid account. NCBI does not charge for the data directly.

That works, thanks a lot. Notice that another optional argument is --gzip compression.

Log in to answer this question.