[gringer]

For a bit more understanding of the different FASTQ formats, have a look at the Wikipedia page:

https://en.wikipedia.org/wiki/Fastq#Encoding

The example sequence in that section looks pretty close to your format.

[dpryan]

Google drive, dropbox, copy.com, there are a few option out there for sharing bigger files. What sort of error do you get when you run bowtie? It has an option to tell it how the phred scores were encoded (and anyway, one could always change them with a simple script).

[ron128]

The mystery Deepens

@ Mr Ryan: Tried what you suggested. Turns out it IS pre 1.8

Tried out solexa QA as well. Returns Casava 1.3 as the pipeline used for generating the data.

Now here is where things start to get intriguing. I managed to run fastqc on this data, which is now telling me in the report that Casava 1.5 was used for generating the data. It is differing with the solexaQA. might be a minor difference between the 1.3 and 1.5 which both the tools are not able to pick up. What intrigues me is that, I still cannot run this stupid data using bowtie (which is format independant) OR Bwa. ANd the data is not corrupt, because I have managed to run fastqc. I am looking at a few things in here. WIll keep everyone posted. Is there a website someone knows where I can host these reads? Gdrive maybe? or is there something NGS specific? I am thinking this is a great dataset for ppl troubleshooting NGS programs and wanting to learn about NGS in general, to get started working upon

[ron128]

Actually, this is somewhere between illumina 1.4 and 1.7, since the multiplex tag is included. I should also mention that it's lane 5, not 1 (it's tile 1). I'm used to seeing the 1101+ tile numbers from the hiseq...

Dear Mr Ryan, Thanks a ton for all your insights. I am trying to run the grep command and see if it is what you are suggesting. I should have some insights by tomorrow upon greping the data, once i get access to my server. thanks a ton! I will post back tomorrow what i get for this data.

[ron128]

http://www.biostars.org/p/911/
may be this will help

Thanks a ton for this! I tried solexaQA as well, but guess what another weird error I have written in to solexaQA mailing list. awaiting a response.

Well I cannot know for certain whether the file is corrupt. My intuition is that it is not corrupt, as i can open the file perfectly fine in a text editor like emacs or vi.

[dpryan]

Actually, this is somewhere between illumina 1.4 and 1.7, since the multiplex tag is included. I should also mention that it's lane 5, not 1 (it's tile 1). I'm used to seeing the 1101+ tile numbers from the hiseq...

[paa6]

Dear Sir/Madam, I have no clue as to what platform was used for sequencing this data. The only thing which has been told to me is that this data is from NIH3T3 cell lines :/ the sequencing company has shut up shop and unfortunately there is no way to be sure about this. I already tried using the phred64 option in prinseq, but I was getting the same error which says that the input file is not in fastq format. Thanks for your thoughts on this

http://www.biostars.org/p/911/
may be this will help

[paa6]

1 is the lane, 2782 the x coordinate, 993 the y-coordinate, etc.. There's no standard for read names, you could name them anything you want (in fact, they could all be the same if you wanted).

hmm I guess that depends on the version of illumina reads..I haven't seen five numerical...so I got confuse!!!
but it means this is fastq format that's for sure..

[dpryan]

1 is the lane, 2782 the x coordinate, 993 the y-coordinate, etc.. There's no standard for read names, you could name them anything you want (in fact, they could all be the same if you wanted).

[paa6]

Ahh sorry for the confusion mam! lol. anways, I wanted to use prinseq with this data, (i prefer it over fastqc because of its DUST low complexity filters as well as the plots which are more intuitive than fastqc). However that is where the trouble started in the first place. Upon running prinseq I got an error that the files were not in fastq format. I tried running dos2unix, since these files have been on a windows system previously and our current server uses unix. Thanks for your suggestion! looks like I am still stuck at phase one of my project.

may possible that file is corrupted.....

I have again gone through ur file and I have found that ur file has every element for being illumina fastq format except there is something is unusual which I missed out at first place...
HWUSI-EAS1642R = this is the instrument name
0000 = flow cell lane
5 = tile no. within the flowcell lane
1 = x co-ordinate
2782 = y co-ordinate
993 = ??? I have no idea what is this for
#GATCAG = index no.
/1 = means single end
i.e. why I said it's a fastq format...but looks like it's not....or something is wrong...my expertise limited to just illumina

[dpryan]

That's an error in prinseq, then. Try some other program (e.g., fastQC). Alternatively, perform the grep command that I suggested earlier.

[ron128]

@ron128 if this is GAIIx data then it most certainly is in the phred+64 format. If you do want to run PRINSEQ then use the "-phred64" option to account for that format.

Dear Sir/Madam, I have no clue as to what platform was used for sequencing this data. The only thing which has been told to me is that this data is from NIH3T3 cell lines :/ the sequencing company has shut up shop and unfortunately there is no way to be sure about this. I already tried using the phred64 option in prinseq, but I was getting the same error which says that the input file is not in fastq format. Thanks for your thoughts on this

[GenoMax]

@ron128 if this is GAIIx data then it most certainly is in the phred+64 format. If you do want to run PRINSEQ then use the "-phred64" option to account for that format.

[ron128]

ahhh no need to be formal by the way...and I am not sir..I am mam...I have analysed ur file and this is fastq file...and use fastqc for analysis of ur sequence before doing anything further....
http://www.bioinformatics.babraham.a...ojects/fastqc/

Ahh sorry for the confusion mam! lol. anways, I wanted to use prinseq with this data, (i prefer it over fastqc because of its DUST low complexity filters as well as the plots which are more intuitive than fastqc). However that is where the trouble started in the first place. Upon running prinseq I got an error that the files were not in fastq format. I tried running dos2unix, since these files have been on a windows system previously and our current server uses unix. Thanks for your suggestion! looks like I am still stuck at phase one of my project.

[dpryan]

"_" is a valid quality score (it's 30 something).