[Bio-software] Re: Consed: combine amplicons of same individual
(by sven_klages at web.de)
Tue Oct 10 02:59:12 EST 2006
*** Marian Thieme (10/09/06 10:49):
> Hello Bastien,
> as you already supposed correctly I have many different individuals.
> Sorry, I didnt point out, that I would like to align each amplicon
> against a reference sequence. Therefore each amplicon can correctly
> placed in the editor/browser. Furthermore I could imagine, that the
> browser combines fragments implicitly, because it only needs to check
> some pattern in the amplicon's filename (for instance) and put all
> amplicons in the same line, where the pattern occurs. Hence we actualy
> dont need to make a big assembly, or not ? and the main task is to align
> all the sequences to a reference sequence, when aligned correctly, then
> the sequence of each individual should assembled implicitly. Make sense ?
I'm not really sure if it is a good strategy to align all amplicons
of a given individual against a reference sequence.
IMHO it is better to align all individuals (or a given subset) of
one amplicon against a reference sequence.
So you have one alignment per amplicon with N individuals.
You're looking for (heterozygous) variations in sequence trace data,
aren't you ?
> (I also tested staden package a little bit, but I found it difficult to
> incorporate a reference sequence, because you first need to create trace
> files in both directions, but I dont think this makes much sense,
> because I get the reference sequence as fasta file without trace data.
You could use 'mktrace' from the Consed Package or 'sudophred' from
PolyPhred, even Staden has a 'makeSCF'.
All tools seem to have problems with large reference sequences.
> Unlike staden, consed is able to do that (with crossmatch))
Or just fasta2Phd.perl/phd2Ace.perl .
You might also take a look at commercial packages like 'SeqScape' from
ABI or 'CodonCode Aligner' among others.
The decision which package to take is IMHO highly dependent on the size
of the project(s).
Working with larger projects works best with Consed/polyphred in our
hands, as many things are can be done via scripts.
> Bastien Chevreux schrieb:
>> Marian Thieme wrote:
>>> I have a question about the Alignment und Sequence editor Consed,
>>> because I want to evaluate and compare some sequence analysis tools to
>>> find out which one is best suited for a resequencing project: Is it
>>> possible to combine fragments/amplicons so that I get a contiguous
>>> sequence. I could imagine, that all fragments of a certain individual
>>> are placed in one line, so I can easly browse and edit the sequence for
>>> each individual. (Perhaps, overlapping fragments should be merged in
>>> some way). (At the moment I havent managed that, so all amplicons are
>>> treated seperatly.)
>> Hi Marian,
>> do you have sequences from one individual or from several? In the later
>> case, I am not sure whether there is a package that has exactly what you
>> want out of the box. In any way, consed (the editor) alone will probably
>> not help you as you would need an assembler (e.g. mira, phrap, cap3, etc.)
>> to go along with it.
>> Have you looked at the Staden package (http://staden.sourceforge.net/)? It's
>> *the* "swiss-army-knife" of assembly tools and is used for very small and
>> very large projects (millions of sequences). It has
>> - a nice preprocessing package (pregap)
>> - a powerful editor (gap4 for contig editing, joining, splitting, annotation
>> and much more)
>> - and a number of very useful tools to complement certain needs (mutation
>> detection etc.).
>> It also has some simple assemblers integrated, external assemblers can be
>> plugged into quite easily.
>> In your case, I could imagine that letting assemblers run on subsets of your
>> sequences (each one containing sequences from exactly one individual)
>> combine the results within gap4 might be what you need.
More information about the Bio-soft