fastq-dump for incomplete files

[Fernas]

Hi All,

I huave a big list of SRA accession numbers (e.g. SRR651693, ..etc). I used fastq-dump to download all these accessions and convert to directly to fastq (without the need to store sra in the disk).
However, I Noticed that for some of the files, the number of reads is different than the number of reads for that file in the NCBI website, which means that the download operation for this dataset was incomplete.

My question is:
is there any validation tool within sra toolkit that I can call (within a script to iterate through fastq files) so it compares the downloaded fastq file with the metadata in the website?
(I know there vdb-validate, but it works only for sra files not for fastq files).

[hdashnow]

I would like to echo this question. Is there some way to download an md5 sum or similar to check the fastq files against?

[piet]

Is there some way to download an md5 sum or similar to check the fastq files against?

'fastq-dump' will first download the SRA file and store it on your local disk. After download is finished, it will convert the SRA file to FASTQ.

If you want to check the integrity of the downloaded SRA file, you can break the procedure into several steps. First download the SRA file with 'prefetch'. Then check its integrity with 'vdb-validate'. And finally extract the reads with 'fastq-dump'.

On the command line you can use the AND operator ('&&') to stop precessing after a failure.

prefetch SRR651693 && vdb-validate SRR651693 && fastq-dump SRR651693

[hdashnow]

'fastq-dump' will first download the SRA file and store it on your local disk. After download is finished, it will convert the SRA file to FASTQ.

If you want to check the integrity of the downloaded SRA file, you can break the procedure into several steps. First download the SRA file with 'prefetch'. Then check its integrity with 'vdb-validate'. And finally extract the reads with 'fastq-dump'.

On the command line you can use the AND operator ('&&') to stop precessing after a failure.

prefetch SRR651693 && vdb-validate SRR651693 && fastq-dump SRR651693

Thanks peit. That worked perfectly.