Illumina solexa 75bp format problem

[anyone1985]

I don't know why every read ends with 22 Ns. Please tell me.

@HWI-EAS241:5:1:10:83#0/1
GCCCCGTCCATCACTTCTGCGATGCCGCGAATGCCCAATGGCAAGCCGNCGGGNNNNNNNNNNNNNNNNNNNNNN
+HWI-EAS241:5:1:10:83#0/1
[a``_`X_O\Q\YQ[Z\O[a\WXNXZZBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBB
@HWI-EAS241:5:1:10:1808#0/1
TGCTGCGGCCCAATGGAGCCACGTTGCCCTGGTGCTTGCCCTTGGGATNGTGGNNNNNNNNNNNNNNNNNNNNNN
+HWI-EAS241:5:1:10:1808#0/1
[aaaaaaa\UX_aaa\U__`a`a`a_^Ua``P\a_aa_\TWa`BBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBB
@HWI-EAS241:5:1:10:1866#0/1
TGGCCGCCTGCGTCACGCCGATTGTCAGCGCCGTGGGCCATGAAACCGNCGTGNNNNNNNNNNNNNNNNNNNNNN

[maubp]

There is something else funny in those records - the spaces in both the sequences and the quality strings. Are those spaces real, or some a cut & paste corruption, or quirk of the forum editor?

[kmcarr]

When you say "every read" do you literally mean EVERY read? Is it the entire flow cell, one lane, part of lane? Did anything happen to the instrument between cycles 53-54, such as reagents being refilled or software restarted?

[anyone1985]

Yes, every read. I used the velvet to assemble the genome. I did not know if it would affect the result of assemble. Whether should I remove the Ns first?

[dcjamison]

The only time a N gets put into the sequence is when the base caller cannot match a cluster in the current tile. Typically this happens at the edge, when clusters "wander" on and off the image. Based on the fact your read quality went kaput in last 20-odd bases, I would guess one of the reagents ran out or was bad -- most likely the incorporation mix -- and you got no cluster illumination.

You do need to trim the N's out before you put the sequences into velvet. Probably easiest to do by rerunning gerald with the USE_BASES param set to Y52n*.

Edit: although it does occur to me the N followed by 4 called bases **might** indicate a laser issue -- highly unlikely, in my opinion, but you might want to discuss it with your FAS.

[maubp]

If you want to just edit the FASTQ file, here is a tiny Biopython script to do this for you (take just the first 52 bases of each read):

from Bio import SeqIO	 
trimmed= (rec[:52] for rec in \	 
          SeqIO.parse(open("original.fastq"), "fastq"))	 
out_handle = open("trimmed.fastq", "w")	 
SeqIO.write(trimmed, out_handle, "fastq")	 
out_handle.close()

That should work on Biopython 1.51 or later (and probably 1.50 from memory).