[mira_talk] Re: 454 and solexa pared-end

Hi Bastien,

hehe, you had busy night to answer all our emails. Thanks 8)

Bastien Chevreux wrote:
> There no such guide :-) I refused until now to think of de-novo assembly with 
> 36mers. Now that SOlexa has 76mers (and with a pretty good quality), I have 
> started work on de-novo hybrid 454/Solexa assembly.
>   
I am really looking forward to.

We are using 36mers for mate-pairs paired-end on solexa, because 76 mers
seems to have too much chimeras in them. For 36mers there is ~1%
chimeras, for 76mers >10%.  From the technology point of view
(preparation of the library) I think you cannot get rid of them. 76mers
are ok for normal paired-end, but then your insert size is ~500, which
would be almost covered by one 454 read and IMHO this kind of paired-end
do not really improve assembly. For longer insert sizes on solexa people
are now stuck with 36mers - until solexa starts to use linker to find
exact position of ligation (like 454) or some other trick.
> Current best alternative so far: assemble the 454 sequences and use these as 
> backbone for a Solexa mapping. Or use Velvet or any other Solexa de-novo 
> program, fragment that data and mix with the 454 reads. Then use that for a 
> mapping with the Solexa data.
>
> All of these approaches already work quite well. Not ideal, but doable.
>   
I tried this, but it looks mira do not use paired-end info from solexa
reads to join appropriate contigs or to split wrong assemblies (which is
probably not possible in mapping).  Am I right or I set it up wrong?
Also when using output from velvet, I loose that valuable paired-end
information.

Well, mira is best, I'll wait for new version. I have real-life data and
I will do beta-testing, if needed.

Thanks,

Jan
 

-- 
You have received this mail because you are subscribed to the mira_talk mailing 
list. For information on how to subscribe or unsubscribe, please visit 
http://www.chevreux.org/mira_mailinglists.html

Other related posts: