Needle and consensus seq?

[Neuls]

Hi,

Our lab is currently sequencing a 1000pb gene fragment. They sequence foward and reverse strand, so they get the same strand sequenced twice. Reverse output is processed in order to get reverse complementary. Then a needlemann alignment is done between forward and reverse complementary. One of the strands from needle output is cuted handly without an standard criteria and this is called 'consensus sequence'.

Can you actually do a consensus from 2 sequence? I guess not. Moreover, I tryed to use cons program from emboss package in order to try get a real consensus sequence from needle because i want to automatize this process. Cons command returns me an error about input file. Looking at its manual it's clearly said the input file must be a multiple sequence alignment, so far I remember, needle is not. Actually, I wonder how you can get a consensus from only 2 sequence, if there is a change in a nucleotide position both situations are represented equaly because theres only two sequences.

Is the process currently done correct?

I got the feeling it would be enought if only one strand got sequenced or if the Blastn was run twice, one for forward and the other one for reverse complementary.

What is your opinion?


Thank you

[dcameron]

I'm not quite sure what your workflow is that you would require a consensus from only a read pair and not be calling a consensus from the totality of reads at a given location but nevertheless, you have a few options:

* Filter out pairs that don't match (not viable for long read technologies since their error rate is so high)
* Call N bases where the strands differ
* Call the sequence with the highest base quality at the position where they differ
* Use a more detailed error model that reflects the sequencing errors found in your data

What are you using to do your sequencing?

[Neuls]

Thank you for replying,

Yes, they require me to get the 'consensus' sequence from only a pair of reads which are the forward strand and the reverse complementary from the same DNA fragment.
I'm not sure that the consensus sequence concept can only include 2 sequence as an input..
The sequening is done by sanger. So far i know i can only extract fasta sequences from these chromatograms.

Can I filter somehow the chromatogram with highest quality?

Bases never differ in the alignment, ethier they are aligned with a N base from the complementary one or a gap is introduced aligning the same kind of bases..