Using variables with fasterq-dump?

Question

I am trying to download multiple fastq files from the SRA NCBI database which, conveniently enough, have their IDs in a range. Building a simple loop doesn't seem to do the trick, as bash appears to interpret my variable as an additional command? Help would be appreciated. This is what I have for now:

for i in {256..345};do
    sra_code = "DRR163""$i"
fasterq-dump "$sra_code" -p -v --outdir ./output/ -e 8 

done
```

Answer 1

You have an obvious error in your shell script that should have given you an error message:

sra_code = "DRR163""$i"

Should be:

sra_code="DRR163""$i"

You cannot have spaces around the = sign in a variable assignment. The line sra_code = "DRR163""$i" means "run the command sra_code with the arguments = "DRR163""$i". Since there is no such command, you see an error.

Here's a working version of your command:

for i in {256..345}; do 
    sra_code="DRR163$i"
    fasterq-dump "$sra_code" -p -v --outdir ./output/ -e 8
done

Answer 2

which, conveniently enough, have their IDs in a range

I think it's a bit risky to rely on that - for this particular case they are not going to change but you may need to download more samples in the future.

I would suggest avoiding sra and fastq(er)-dump altogether and use ena instead.

In your case, it seems you want the runs in project PRJDB7736. From here https://www.ebi.ac.uk/ena/browser/view/PRJDB7736 you can get a tabular file (see link in Download report ... TSV) with the sample characteristics. Column fastq_ftp gives you the ftp link to the fastq files that you can pass to wget or curl.

You can also get the same report programmatically with:

curl -s -X GET "https://www.ebi.ac.uk/ena/portal/api/filereport?accession=PRJDB7736&fields=run_accession,sample_description,library_name,library_layout,fastq_ftp&result=read_run" > ena.tsv

You should edit the fields=... list to suite your needs. The fields available are those from:

curl -X GET "https://www.ebi.ac.uk/ena/portal/api/returnFields?dataPortal=ena&result=read_run"

I don't know what you think... I find this route much more useful and transparent than sra's.