[mira_talk] Re: Problem with convert_project
- From: Bastien Chevreux <bach@xxxxxxxxxxxx>
- To: mira_talk@xxxxxxxxxxxxx
- Date: Sun, 28 Oct 2012 09:43:56 +0100
On 10/26/2012 09:14 PM, Steven Sijmons wrote:
[...]
When I then compared the number of reads in the fastq file I created
with convert_project, I noticed that these contained more reads than
the number of lines present in the debris_list. Debris_lists contained
232 628, 2 288 887 and 42 826 127 reads, while the convert_project
outputs contained 437 106, 2 392 084 and 43 632 082 reads
respectively. Could these 'extra' reads be causing this? And where are
these reads coming from if they aren't in the debris_list? Hoping that
someone can give some possible explanations for this, because I really
don't understand what's happening here.
Thanks for the data. Am I correct in guessing you used MIRA 3.4.x? Well,
I'm quite sure of it actually.
Two things play a role in this misbehaviour you are seeing:
- MIRA 3.4.x does not recognise CASAVA 1.8 style Illumina paired-end
data. One needs to rename reads before 3.4.x sees that they are paired
end. You did not do that.
This naming convention problem also applies to convert_project from
the same package.
- MIRA 3.4.x has a broken "check for doubly appearing reads" routine ...
it does not recognise when two reads with the same name are given. This
happens with CASAVA 1.8 style naming.
Possible fixes are simple, but you will need to redo assemblies:
1) if you want to continue with MIRA 3.4.x, you need to rename Illumina
reads from CASAVA pipelines >= 1.8. That is, if a FASTQ definition is
@somename 1:whatever
you need to rename it to
@somename/1 1:whatever
and similarly for the other half of the pair ("2" instead of "1" above)
2) You switch to MIRA 3.9.x and the naming problems are automatically solved
Hope that helps,
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:
