[mira_talk] Re: Metagenomic 454 Titanium Assembly
- From: Bastien Chevreux <bach@xxxxxxxxxxxx>
- To: mira_talk@xxxxxxxxxxxxx
- Date: Thu, 14 Jan 2010 19:56:49 +0100
On Mittwoch 13 Januar 2010 Ryan Lesniewski wrote:
> First, I want to say that MIRA is an incredibly flexible, and
> well-documented/explained program....big thanks! It is incredibly hard to
> come by this rare breed these days.
Thanks.
> Now, my big question is: Is MIRA flexible enough to handle about 2 million
> (~350 bp avg.) metagenomic reads from an environmental sample?
2m 454 reads should be no big problem.
> Of course,
> this dataset has many different organisms. Can I strap up the quality and
> stringency so that I can get organism specific, non-chimeric contigs? Do
> you have recommendations for parameter adjustments?
Hmmm, never did that myself. After giving it a first thought, here's what I
would try first:
--job=denovo,genome,accurate,454 -GE:urd=no
as basis. You will very probably still get some contigs with reads from
several organisms which differ by only a few bases but will low coverage.
You can try to adjust sensitivity down a bit
454_SETTINGS -CO:mrpg=3
or even mrpg=2, but this will also create "false positive" repeat detections
and these lead to additional contigs.
Also play a bit around with
454_SETTINGS -AL:mrs=90
to increase SW alignment stringency.
Last but not least: the "proposed end clipping" will be on by default. This
may be a good thing for you as this will clip back or throw out single reads
which are quite different from others, ie.e, were sequence from rare
organisms. You might want that. If you prefer have contigs a tad longer, try
switching off "-CL:pec" but beware that this function is also a very powerful
error detection mechanism.
> I also have metatranscript(cDNA) datasets from the same environments. Any
> suggestions on using the EST function to assemble some of the transcripts?
Just fill them in and see what happens. I very much hope there won't be too
many RNA genes expressed.
Regards,
Bastien
--
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:
