SRA to fastq conversion with fastq-dump loses sequences

[pcantalupo]

Hello,

I converted an SRA archive (ftp://ftp-trace.ncbi.nlm.nih.gov/sra...953/SRR073769/) to fastq with the fastq-dump program (sratoolkit-2.1.6). The resulting fastq file had ~160,000 less sequences (2% of the total number of spots) than expected. Why does this occur?

Thank you,

Paul

[eeh_021]

I've also experienced this problem. Did you find a solution?

Thank you,

Elizabeth

[ymc]

how do u know the expected number of sequences?

[GenoMax]

I am seeing the same number of sequences as reported on the SRA page:

http://www.ncbi.nlm.nih.gov/sra?term=SRR073769

in the file I downloaded.

../sratoolkit.2.1.16-centos_linux64/bin/fastq-dump.2.1.18 SRR073769.sra 
Written 8175900 spots for SRR073769.sra
Written 8175900 spots total

$ more SRR073769.fastq | grep "@SRR073769" | wc -l
8175900

[eeh_021]

Oh, I see.

I'm still a little confused about something:

This file, SRR035116.sra, for example, is 3.9Gb
When I convert it to fastq, however, it is only 2.2Gb. I checked the number of spots, and it's, surprisingly, the same.

Usually when I convert sra to fastq, my files get a lot bigger. Help?

Thank you!!

[GenoMax]

.sra file is ~383 Mb and the .fastq file is 1.6 G (on my filesystem). If your .sra file is truly that large then something must be wrong.

Use the aspera client that SRA provides to download the .sra file.

I'm still a little confused about something:

This file, SRR035116.sra, for example, is 3.9Gb
When I convert it to fastq, however, it is only 2.2Gb. I checked the number of spots, and it's, surprisingly, the same.

Usually when I convert sra to fastq, my files get a lot bigger. Help?

Thank you!!

[eeh_021]

383Mb?

http://www.ncbi.nlm.nih.gov/sra?term=SRR035116

If you go there, it is supposed to be 3.9Gb, and that's about how big it is when I download it...

[seb567]

383Mb?

http://www.ncbi.nlm.nih.gov/sra?term=SRR035116

If you go there, it is supposed to be 3.9Gb, and that's about how big it is when I download it...

Perhaps it is more straightforward to fetch it from Europe or Japan.

Compressed files (.fastq.gz or .fastq.bz2) are just easier to use than those .sra files.


Sébastien Boisvert

[GenoMax]

We are talking about two different data sets.

My response was for the dataset (SRR073769) that was in pcantalupo's original post.

Dataset you are referring to below is indeed 3.9 Gb.

383Mb?

http://www.ncbi.nlm.nih.gov/sra?term=SRR035116

If you go there, it is supposed to be 3.9Gb, and that's about how big it is when I download it...

[csmatyi]

Oh, I see.

I'm still a little confused about something:

This file, SRR035116.sra, for example, is 3.9Gb
When I convert it to fastq, however, it is only 2.2Gb. I checked the number of spots, and it's, surprisingly, the same.

Usually when I convert sra to fastq, my files get a lot bigger. Help?

Thank you!!

I don't think that's a problem if the fastq file gets bigger because the sra file is in binary anyway, which is more compact.

[alireda82]

how to convert SRA file to FASTQ?

[GenoMax]

how to convert SRA file to FASTQ?

Use SRA toolkit: http://eutils.ncbi.nih.gov/Traces/sr...lkit_doc&f=std

[VC87]

Hi everyone, i'm new here!
Can someone tell-me if it's possible to cenvert a WIG file type to FASTQ?thanks in advance

[blancha]

No, the Wig files do not contain the sequences, just the coverage.