Cell Ranger
Why NCBI SRA _1/_2 FASTQ names need review before Cell Ranger
fastq-dump writes SRR…_1.fastq.gz. Cell Ranger wants an Illumina suffix. The mapping is usually R1/R2, but three-file dumps exist, so the planned renamer proposes a name and does not apply it until you confirm read identity.
What NCBI actually wrote
fastq-dump --split-files names mates by a trailing _1 / _2. A typical pair is SRR9291388_1.fastq.gz and SRR9291388_2.fastq.gz. cellranger count does not look for that layout. It looks for an Illumina suffix: [Sample]_S#_L###_R#_001.fastq.gz.
10x’s SRA article uses that same SRR9291388 example and maps _1 → R1 and _2 → R2, then inserts _S1_L001_001. That is the proposal the downloadable will print.
Why it is not a silent rename
The mapping is an inference about read identity, not a character substitution. Two-file dumps usually are R1/R2. Three-file dumps (index plus barcode/UMI plus cDNA) often mean _1 is I1, _2 is R1, and _3 is R2. Renaming the first file to R1 in that case points Cell Ranger at the wrong read.
So the program marks sra_read_layout as needs-review and does not apply the rename until you confirm. Confirm chemistry and read lengths before you accept it. We do not open the FASTQ to guess for you.
What to do with the proposal
If you have a two-file 10x SRA set and the read lengths match barcode/UMI versus cDNA, the proposed names are the ones Cell Ranger will accept. When the downloadable ships, accept those two renames and run --sample=SRR9291388 (or whatever prefix you kept). If a third file is present, stop and assign I1/R1/R2 from the lengths, not from our default.