A community effort to collect a curated set of analysis pipelines built using Nextflow.

nf-core

Hi, currently <https://nf-co.re/modules/fastp/|FASTP> only accepts gziped FASTQ files (<https://github.com/nf-core/modules/blob/b25dc591c1add843d9a8b03cdcfba5e90675e396/modules/nf-core/fastp/main.nf#L84|e.g. here>). Is there a reason for this? Asking because the combination <https://github.com/ncbi/sra-tools/wiki/08.-prefetch-and-fasterq-dump|prefetch + fasterq-dump> from NCBI’s SRA tools only produces uncompressed FASTQ files.

For reference, the <https://nf-co.re/modules/sratools_fasterqdump/|fasterq-dump> module compresses the files <https://github.com/nf-core/modules/blob/73da03cd2508eda722a32111113e0d7da353abcd/modules/nf-core/sratools/fasterqdump/main.nf#L48|here>.

Not sure about the module, <https://github.com/OpenGene/fastp?tab=readme-ov-file#input-from-stdin|but the tool itself supports uncompressed input> - at least from `STDIN`.

Ultimately, `gzip` will bring down a FASTQ to about 20% of its size, but at a high computational cost.

For data at rest, that is perfectly sane (_even though much better archival formats exist_), but for an analysis pipeline, it is actually wasteful to go through repeated cycles of de- and recompression just for trimming, UMI reformatting and such. Fastp at least can do all of this in one pass.