How to convert sra-lite format to fastq?

[tbusch0000]

I am trying to dump sra-lite (sequence read archive) files to fastq format. On the NCBI Sequence Read Archive site it states:

...users are asked download runs of interest and execute dumps into the desired format using the SRA SDK toolkit available at http://www.ncbi.nlm.nih.gov/Traces/s...are&s=software

I downloaded the precompiled toolkit for 64-bit architecture onto my macbookpro running snow leopard and tried to run the fastq-dump executable from the terminal, and get the error message "cannot execute binary file".

Any guidance would be much appreciated!

[SongLi]

Although I can get their CentOS 64bit running, it's realy slow, take about 10hrs to unpack one file. I am also interested to know more about this new SRA-tools.

[tbusch0000]

I just noticed they released a new MacOSX beta package.

I downloaded that one and entered in the terminal $./fastq-dump -A SRP000910 -D SRR070499.lite.sra

Received error message: "memory exhausted while constructing memory map within file system module - failed to open 'SRR070499.lite.sra'"

[maubp]

Originally Posted by tbusch0000

I downloaded the precompiled toolkit for 64-bit architecture onto my macbookpro running snow leopard and tried to run the fastq-dump executable from the terminal, and get the error message "cannot execute binary file".

My guess is you download a 64bit Linux binary, which won't work on the Mac.

[tbusch0000]

Originally Posted by maubp

My guess is you download a 64bit Linux binary, which won't work on the Mac.

Thanks, they've only just released the mac binaries. It will execute now, but gives the error message above.

[maubp]

Originally Posted by tbusch0000

Received error message: "memory exhausted while constructing memory map within file system module - failed to open 'SRR070499.lite.sra'"

How much RAM do you have, and how big is SRR070499.lite.sra?

[tbusch0000]

Originally Posted by maubp

How much RAM do you have, and how big is SRR070499.lite.sra?

I have 6GB RAM and the file is 3.5 GB

[seb567]

I have to download and convert files to test Ray, the assembler I am working on (see a thread elsewhere on this forum).

My take on sratoolkit (I use /software/sratoolkit.2.0b4-2-centos_linux64/):

It is slow, but it works. My guess is that data are compressed, using something like LIBBZ2 (it is just a guess). That explains the compression ratio as well as the slowness.

[boiseb01@ls30 MyShortReadArchive]$ ldd /software/sratoolkit.2.0b4-2-centos_linux64/fastq-dump
linux-vdso.so.1 => (0x00007fff361ff000)
libdl.so.2 => /lib64/libdl.so.2 (0x00000033f5a00000)
libz.so.1 => /lib64/libz.so.1 (0x00000033f6600000)
libbz2.so.1 => /lib64/libbz2.so.1 (0x0000003403e00000)
libm.so.6 => /lib64/libm.so.6 (0x00000033f5600000)
libc.so.6 => /lib64/libc.so.6 (0x00000033f5200000)
/lib64/ld-linux-x86-64.so.2 (0x00000033f4e00000)

Binaries are linked against libz and libbz2, but the slowness indicates that they probably rely on libbz2.

[maubp]

I'm not 100% sure how memmap works on the Mac, but it sounds like you should have enough RAM to load the whole file into memory (assuming no other memory hungry applications are running at the same time). Can you find a smaller example to test?

[SongLi]

Hi seb567,

How slow are you experiencing with fasta-dump?

My experiene is this: my computer is Xeon 2.4G 4core, 12G RAM, fasta-dump takes 600 minutes to finish one sra file.

I have tried the newest release and also different sra files. fastq-dump is always very slow.

Thanks,

Originally Posted by seb567

I have to download and convert files to test Ray, the assembler I am working on (see a thread elsewhere on this forum).

My take on sratoolkit (I use /software/sratoolkit.2.0b4-2-centos_linux64/):

It is slow, but it works. My guess is that data are compressed, using something like LIBBZ2 (it is just a guess). That explains the compression ratio as well as the slowness.



Binaries are linked against libz and libbz2, but the slowness indicates that they probably rely on libbz2.

[seb567]

About 1-2 hours for a 2 GB sra file, though it is very approximated.

I downloaded all sra files for SRA010766, converted them from sra to fastq, then to fastq.gz. The script started yesterday 6 PM (EST).

So yours is slower, way slower.

[boiseb01@ls30 Illumina-SRX015621]$ ls
batch-3 SRR033559_1.fastq.gz SRR033570_1.fastq.gz SRR033581_1.fastq.gz SRR033592_1.fastq.gz SRR033603_1.fastq.gz SRR033614_1.fastq.gz SRR033625_1.fastq.gz
download.log SRR033559_2.fastq.gz SRR033570_2.fastq.gz SRR033581_2.fastq.gz SRR033592_2.fastq.gz SRR033603_2.fastq.gz SRR033614_2.fastq.gz SRR033625_2.fastq.gz
files.txt SRR033560_1.fastq.gz SRR033571_1.fastq.gz SRR033582_1.fastq.gz SRR033593_1.fastq.gz SRR033604_1.fastq.gz SRR033615_1.fastq.gz SRR033626_1.fastq.gz
list-sra.sh SRR033560_2.fastq.gz SRR033571_2.fastq.gz SRR033582_2.fastq.gz SRR033593_2.fastq.gz SRR033604_2.fastq.gz SRR033615_2.fastq.gz SRR033626_2.fastq.gz
newFiles SRR033561_1.fastq.gz SRR033572_1.fastq.gz SRR033583_1.fastq.gz SRR033594_1.fastq.gz SRR033605_1.fastq.gz SRR033616_1.fastq.gz SRR033627_1.fastq.gz
nohup.out SRR033561_2.fastq.gz SRR033572_2.fastq.gz SRR033583_2.fastq.gz SRR033594_2.fastq.gz SRR033605_2.fastq.gz SRR033616_2.fastq.gz SRR033627_2.fastq.gz
README SRR033562_1.fastq.gz SRR033573_1.fastq.gz SRR033584_1.fastq.gz SRR033595_1.fastq.gz SRR033606_1.fastq.gz SRR033617_1.fastq.gz SRR033628_1.fastq
SRA010766 SRR033562_2.fastq.gz SRR033573_2.fastq.gz SRR033584_2.fastq.gz SRR033595_2.fastq.gz SRR033606_2.fastq.gz SRR033617_2.fastq.gz SRR033628_2.fastq
SRR033552_1.fastq.gz SRR033563_1.fastq.gz SRR033574_1.fastq.gz SRR033585_1.fastq.gz SRR033596_1.fastq.gz SRR033607_1.fastq.gz SRR033618_1.fastq.gz SRR033629_1.fastq
SRR033552_2.fastq.gz SRR033563_2.fastq.gz SRR033574_2.fastq.gz SRR033585_2.fastq.gz SRR033596_2.fastq.gz SRR033607_2.fastq.gz SRR033618_2.fastq.gz SRR033629_2.fastq
SRR033553_1.fastq.gz SRR033564_1.fastq.gz SRR033575_1.fastq.gz SRR033586_1.fastq.gz SRR033597_1.fastq.gz SRR033608_1.fastq.gz SRR033619_1.fastq.gz SRR033630_1.fastq
SRR033553_2.fastq.gz SRR033564_2.fastq.gz SRR033575_2.fastq.gz SRR033586_2.fastq.gz SRR033597_2.fastq.gz SRR033608_2.fastq.gz SRR033619_2.fastq.gz SRR033630_2.fastq
SRR033554_1.fastq.gz SRR033565_1.fastq.gz SRR033576_1.fastq.gz SRR033587_1.fastq.gz SRR033598_1.fastq.gz SRR033609_1.fastq.gz SRR033620_1.fastq.gz SRR033631_1.fastq
SRR033554_2.fastq.gz SRR033565_2.fastq.gz SRR033576_2.fastq.gz SRR033587_2.fastq.gz SRR033598_2.fastq.gz SRR033609_2.fastq.gz SRR033620_2.fastq.gz SRR033631_2.fastq
SRR033555_1.fastq.gz SRR033566_1.fastq.gz SRR033577_1.fastq.gz SRR033588_1.fastq.gz SRR033599_1.fastq.gz SRR033610_1.fastq.gz SRR033621_1.fastq.gz SRR033632_1.fastq
SRR033555_2.fastq.gz SRR033566_2.fastq.gz SRR033577_2.fastq.gz SRR033588_2.fastq.gz SRR033599_2.fastq.gz SRR033610_2.fastq.gz SRR033621_2.fastq.gz SRR033632_2.fastq
SRR033556_1.fastq.gz SRR033567_1.fastq.gz SRR033578_1.fastq.gz SRR033589_1.fastq.gz SRR033600_1.fastq.gz SRR033611_1.fastq.gz SRR033622_1.fastq.gz SRR033633_1.fastq
SRR033556_2.fastq.gz SRR033567_2.fastq.gz SRR033578_2.fastq.gz SRR033589_2.fastq.gz SRR033600_2.fastq.gz SRR033611_2.fastq.gz SRR033622_2.fastq.gz SRR033633_2.fastq
SRR033557_1.fastq.gz SRR033568_1.fastq.gz SRR033579_1.fastq.gz SRR033590_1.fastq.gz SRR033601_1.fastq.gz SRR033612_1.fastq.gz SRR033623_1.fastq.gz
SRR033557_2.fastq.gz SRR033568_2.fastq.gz SRR033579_2.fastq.gz SRR033590_2.fastq.gz SRR033601_2.fastq.gz SRR033612_2.fastq.gz SRR033623_2.fastq.gz
SRR033558_1.fastq.gz SRR033569_1.fastq.gz SRR033580_1.fastq.gz SRR033591_1.fastq.gz SRR033602_1.fastq.gz SRR033613_1.fastq.gz SRR033624_1.fastq.gz
SRR033558_2.fastq.gz SRR033569_2.fastq.gz SRR033580_2.fastq.gz SRR033591_2.fastq.gz SRR033602_2.fastq.gz SRR033613_2.fastq.gz SRR033624_2.fastq.gz

[tbusch0000]

Thanks for the tips.

I got the fastq-dump working on an x-large amazon cloud instance running cent os ami.

[babaref]

How to convert fastq format to sra files? is there any perl script for this conversion?

[gsgs]

I want the table, that converts a byte from the sra file
into a sequence of nucleotides

http://www.flutrackers.com/forum/showpost.php?p=507401

SRA toolkit sourcecode has "4na" and "2na"

[dpryan]

Why don't you either use fastq-dump or just download the gzipped fastq files from ENA (such as this one)?

[gsgs]

I found that other thread, saying that the format is complicated,
so there is no such table.

http://seqanswers.com/forums/archive...p/t-12054.html


I'm having problems with files >4GB and wanted to test it
on a partially downloaded file first

I have to split the large files, so they work with my programs.
It's also faster, better for testing, dealing with 4GB files is tedious.

I doubt that sra-tools will work with such splitted files

[gsgs]

OK, I tried to download the file to my external drive, it took 5.5h ,
until an error message was displayed that the file couldn't be copied.

Then I searched my main HD and found that it was put into a temporary file
which had 4631463048 Bytes, so apparently >4GB is possible on my
main drive but not on the external one.
(Windows XP, computer bought in 2010 or 2011)


I made a copy of that temporary file to another file on the maindrive,
then I closed the error window, and indeed, the temporary file was
deleted, but luckily I had the copy.
As expected I can't copy that file to the external drive nor can I access it
with any of my programs.
But DOS-commands copy,type,find do work.

So, I need a program that splits the big file into 2 smaller files, that can be assessed.

[GenoMax]

If you are using 32-bit windows XP (which you likely are) this may not be possible. What kind of format do you have on your external drive? You may need NTFS for files > 4GB.

[dpryan]

Ah, yeah, I expect that the SRA format is pretty non-trivial from the various discussions of it. Honestly, if your computer is having issues with files ~4GB then you might just be better off using someone else's (though check if the drive is NTFS formatted), particularly if you're stuck on windows. Got a labmate with a Mac?

[gsgs]

I cannot just switch to Win64, since I need all my old programs
that were written on 16bit or 32bit