SRAtoolkit error during fastq-dump

Question

[dally]

I downloaded an SRA file from GEO FTP. I then ran the following command trying to grab the fastq files, but I keep coming up with this error and I can't find anything about it on google.

Here's the command used:

fastq-dump --split-3 '/path/to/SRR1460694.sra' 

Error:

fastq-dump.2.5.7 err: connection inconsistent while validating within network system module - failed

I am behind a network firewall, but I have my proxy set up in by .bash_profile so I don't usually ever run into problems downloading, or updating software. I'm not sure what to do at this point. Any ideas?

[GenoMax]

If you got the data from EBI then why bother with SRA files. Just get the fastq files directly from ENA-EBI. Links for them should be on the page you get to after you search for the SRA # you need.

[dally]

Sorry, EBI was incorrect, I got the SRA files from GEO.

[GenoMax]

Then get the fastq from ENA-EBI.

[dally]

They sadly do not appear to have any fastq files for download.

Answer 1

[piet]

If you have to use a HTTP proxy, you have to configure it explicitly for sratools. Sratools do not care about standard unix environment variables like http_proxy. NCBI programs have their very own ways of configuration and usage, which often resembles standards from the previous millennium when Fortran programs were fed into computers with punch cards. It seems as if NCBI is still developing with PDP11 in mind.

To make sratools use a HTTP proxy, append the following to your local configuration file which is found at ~/.ncbi/user-settings.mkfg:

/http/proxy/enabled = "true"
/http/proxy/path = "proxy.example.com:8080"

See also http://seqanswers.com/forums/showpost.php?p=179416&postcount=8

Support for HTTP proxies has been added to sratools in 2015. Therefore the above settings will not work with some older versions.

It seems that you have already downloaded the SRA file to your local disk. But fastq-dump still needs access the internet in order to download the related reference sequence. Unfortunately, most SRA files are very similar to BAM files, they comprise a mapping of reads to a reference sequence. The reference sequence is not included in the SRA file. SRA files are not self-contained, you need additional external information to read their content.

[dally]

This actually ended up being the answer. Just had to add my proxy settings to the configuration file. Thanks a ton piet.

[piet]

Yeah, nowadays it's playing quite well with proxies, but still a bit under-documented. One year ago I had to use strace to see where it got stuck, and then wrap it with proxychains. SRA tools are using libcurl under the hood. But they somehow prevent that libcurl can see the environment variable http_proxy.

[cedric.laczny]

Thx! Upvoted. On that note, I just stumbled over http://ncbi.github.io/sra-tools/install_config.html -> vdb-config provides a "graphical" configuration screen. N.B. As indicated by @piet, do not use "http://" in front of the proxy as it won't work then (at least it didn't in my case).