Forum:Download manager for SRA

Question

[Biojl]

Hi,

I was wondering what people is using to download huge amounts of experiments in SRA format. I'm currently using a home made script that downloads everything and classifies it using the runinfo table that can be downloaded from the sra selector (http://www.ncbi.nlm.nih.gov/Traces/study/?go=home)

I find this extremely cumbersome but is the best option I got. Is there any download manager I'm missing? What do you people use to download a (big) set of experiments.

Answer 1

[Istvan Albert]

Well I start with something like this

esearch \
  -db sra \
  -query PRJNA40075 | \
  efetch \
  --format runinfo | \
  cut -d ',' -f 1 | \
  grep SRR | \
  head -5 | \
  xargs fastq-dump -X 10 --split-files

Also seen here: How to download raw sequence data from GEO/SRA

[Zev.Kronenberg]

If you plan to use BWA mem make sure to set the --origfmt flag for the fasts-dump Aligning paired end fastq files dumped from SRA

[Istvan Albert]

never ran into this myself - though I do download/process paired end reads - there is probably more to it, perhaps the way the sra file was built.

Answer 2

[dariober]

Maybe the SRAdb package from R/Bioconductor can help? As mentioned in, e.g. this A: SRX from SRR IDs . It will help deciding what needs to be downloaded.