sra-toolkit not working

Question

[Naren]

I downloaded sra-toolkit from sra website for 64bit Windows.

After extracting I tested it from bin folder:

sra-toolkit\bin> fastq-dump --stdout -X 2 SRR390728

but it throws error:

2015-12-15T09:54:40 fastq-dump.2.5.5 err: item not found while constructing within virtual database module - the path 'SRR390728' cannot be opened as database or table

What may be the cause and how to get it working?

[piet]

I can start the download from my current site. The file size is about 950 Mb. Thus the item is available from NCBI. The two major problem I have faced with SRA toolkit are:

1. if you need a proxy to access the internet, you have to configure this explicitely for SRA toolkit
2. make sure, that you have enough disk space in the cache directory used by SRA toolkit. By default, the cache directory is placed in your HOME directory. If your HOME is mounted via NFS or is otherwise restricted in size, the default is a very bad choice.

Answer 1

[Istvan Albert]

Occasionally it needs this

vdb-config --restore-defaults

[Naren]

The solution works for others but in my case no effect.

[Matt Shirley]

Their issue tracker is here: https://github.com/ncbi/sra-tools/issues

[Istvan Albert]

post on their github as an issue. As with many other bioinformatics software the sratookit produces incredibly user-unfriendly error messages even when the problems may be simple - I got this cryptic error myself, one of these typically fixes it,delete the cache directory, resinstall the whole thing etc.

[Arnaud Ceol]

This worked also for me to resolve another similar problem reported at A: Sra to fastq conversion err :

err: name not found while resolving tree within virtual file system module

Answer 2

[Devon Ryan]

Just download the fastq file from ENA. When possible, make your life simpler and avoid SRA files :)

[Istvan Albert]

the problem there is that ENA can be excruciatingly slow, sratoolkit is a genius idea in that it allows us to download just a subset of the data, which often is sufficient to verify a statement or check a result - alas it is not implemented the right way and the error messages are terrifyinly inefficient - what we really need is a tool to download fastq/bam files partially or download just 1% of the data.

[piet]

SRA toolkit always downloads the whole SRA file to your local disk and stores it in a hidden directory. After download has completed, you may extract only part of the data from your local copy.

[Istvan Albert]

whoa! indeed. I never realized this, my a whole SRA world as I know it came crashing down. In my demos it is always so fast, but then of course because at some point in life I did indeed download these files ...

[Devon Ryan]

On this side of the pond ENA is quite quick :)

But yes in an ideal world SRA would be the better route.