[mira_talk] Re: MIRA RNASeq de novo assmebly using paired-end strand-specific data

  • From: Elton Vasconcelos <eltonjrv@xxxxxxxxx>
  • To: mira_talk@xxxxxxxxxxxxx
  • Date: Tue, 4 Aug 2015 14:02:35 -0300

Thanks Bastien,

I split the very long reads (>30 kb) into smaller ones (20kn with 3kb
overlap, like you suggested) and I think it's gonna work now, cause mira
went to another step of error messages, claiming that some reads have equal
names (it is regarding the split ones). I corrected that and I'm trying
again at this moment. Let's see what happens.

Could you also help me with another error message I'm getting running mira
onto a cluster?
Unlike the little server at my lab, in the cluster (that uses "slurm" job
management system) mira stops while loading the fastq files (during the
zlib warning message).
################################################
[erv3@login3 assembly-1stTrial]$ tail slurm-42733.out

Tmp directory is not on a NFS mount, good.

Localtime: Mon Aug 3 16:44:27 2015

Loading reads from ../all_R1-trimmed_paired.fastq type fastq
Localtime: Mon Aug 3 16:44:27 2015
Loading data from FASTQ file: ../all_R1-trimmed_paired.fastq
(sorry, no progress indicator for that, possible only with zlib >=1.34)
Failure, wrapped MIRA process aborted.
################################################

My fastq files are not compressed; Any clue about why mira on the cluster
is stopping at this point?

Thanks a lot for your support,
Best,
Elton

2015-07-31 18:40 GMT-03:00 Bastien Chevreux <bach@xxxxxxxxxxxx>:

On 31 Jul 2015, at 16:42 , Diego Mauricio Riaño-Pachón <
diego.riano@xxxxxxxxxxxxxxxx> wrote:
I tried the switch -SK:acrc=n, that doe snot really work. The result
will have transcript assembled from only R1 or only R2 read. With the
Illumina dUTP protocol the contigs from R2 reads will be in the correct
orientation, but MIRA would have ignored all the paired information.

Ah, see? No one really told me that until now, and I never came around
working with stranded data.

I think it would be great if eventually MIRA could produce real
strand-specific transcript assemblies.

This is indeed not implemented yet in MIRA. I could think of 2 possible
solutions:

1. have MIRA start contigs with a R2 read. Pro: comparatively easy to
implement. Con: what about the error rate? R2 reads in the wrong direction
may wreak havoc.
2. have MIRA build a contig and then reverse it if "majority" of R2 points
backwards. Pro: ???. Con: routines for doing a contig reversal in MIRA do
not exist atm While not exactly rocket science, there’s a lot to write and
test. Not going to happen soon.

B.


--
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




--
Elton Vasconcelos, DVM, PhD
Post-doc at Verjovski-Almeida Lab
Department of Biochemistry - Institute of Chemistry
University of Sao Paulo, Brazil

Other related posts: