454 Paired End orientation problems

[pr0t3us]

Hi there,

I have a problem with 454 paired-end in a re-sequencing data set.
Newbler tell me that the paired end are in opposite strand (-> <- or <- ->) but when I try to allign (with blastn) the PE reads against the reference I always find that the two arms are on the same orientation (-> -> or <- <-).

Has someone noted the same problem ?
Is this a bug ?

Thanks in advance

Alessandro

[maubp]

Are you using an SFF file with Newbler? If so the reads would be held with the same orientation (read, pair linker, read).

My next question would then be how are you extracting the read pairs from the SFF file to give them to BLAST? Maybe that step is reverse complementing one of the reads...

[pr0t3us]

Hi Maubp,

thanks for your answer!

Are you using an SFF file with Newbler? If so the reads would be held with the same orientation (read, pair linker, read).

I'm using sff files, but when I add these files to newbler and I run an assembly or a mapping the two arms of paired end are aligned on opposite strand but when I align the same read against the reference with BLAST (or BLAT) the two arms are on the same strand.

My next question would then be how are you extracting the read pairs from the SFF file to give them to BLAST? Maybe that step is reverse complementing one of the reads...

I give the entire read. Blast aligns the reads creating 2 hits for the two arms (the stuffer is not aligned) and they are aligned on the same strand!!

[maubp]

I'm using sff files, but when I add these files to newbler and I run an assembly or a mapping the two arms of paired end are aligned on opposite strand but when I align the same read against the reference with BLAST (or BLAT) the two arms are on the same strand.

How are you looking at the assembly or mapping? Are you looking at the ACE file?

I give the entire read. Blast aligns the reads creating 2 hits for the two arms (the stuffer is not aligned) and they are aligned on the same strand!!

Good - in a proper 454 paired end read in the original form from the SFF file then both reads are on the same strand.

[pr0t3us]

How are you looking at the assembly or mapping? Are you looking at the ACE file?

I have the same effect in the assembly and in the mapping.
In the assembly I look at 454PairStatus.txt.
For example this is a line:
J4Q8OU01DGX18 SameContig 8089 contig00092 99079 + contig00092 107168 -

You can see that the paired end reads is mapped on the same contig but in opposite strand

In the mapping I look at the same file (454PairStatus.txt) anf this is a line:
FJ4Q8OU01COB5I TruePair 7324 AP009048.1 2006538 + AP009048.1 2013862 -


But, when I blast this reads against the contig or the reference I obtain the two arms on the same strand....

[flxlex]

Good - in a proper 454 paired end read in the original form from the SFF file then both reads are on the same strand.

maubp is right here. That settles the discussion, doesn't it? Just draw it out! Newbler just converts the mapping information back into the -> <- orientation for clarity.