MIRA output for Illumina mapping giving 100% coverage!

Kasycas · 09-10-2010, 03:52 AM

Hi all,

I've been using MIRA as I have both Illumina and 454 reads that I wanted to do mapping/assembly as well as semi-hybrid and hybrid assembly. However, when I use MIRA on the Illumina data alone for mapping with

"mira --project=0800 --job=mapping,genome,normal,solexa --fastq -SB:bft=gbf >&0800_Illumina_log_assembly.txt"

I get 1 contig of length 4,798,671 which is pretty much 100% coverage of the genome! I've used bwa and bowtie as well which give me a more realistic ~68-69% coverage. Can anyone tell me why/how MIRA is doing this? I've played around with the setting and read the documentation but I'm at a loss!

Thanks,

Kasycas

maubp · 09-10-2010, 04:45 AM

I'd ask on the MIRA mailing list, but Bastien also posts here.

My guess is different rules/settings for mismatch thresholds, but also MIRA isn't really doing a mapping but a guided assembly (it will extend the reference sequence(s) if it can).

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Using MIRA assembler for illumina sequence	robelb4	Bioinformatics	3	06-24-2013 12:33 PM
Dindel giving error for every candidate indel	gaffa	Bioinformatics	4	03-14-2011 12:18 PM
What does mean 2 x 35 or 2 x 100 in illumina solexa?	Stock	General	1	02-14-2010 03:09 AM
Discussion about MIRA hybrid assembly of 454 reads with Illumina unpaired data	edge	De novo discovery	5	11-16-2009 02:17 AM
MIRA:454 scaffolds polishing by Illumina reads discussion	Patrick	454 Pyrosequencing	0	10-16-2009 07:46 PM

09-10-2010, 03:52 AM	#1
Kasycas Member Location: Dublin, Ireland Join Date: Sep 2009 Posts: 22	MIRA output for Illumina mapping giving 100% coverage! Hi all, I've been using MIRA as I have both Illumina and 454 reads that I wanted to do mapping/assembly as well as semi-hybrid and hybrid assembly. However, when I use MIRA on the Illumina data alone for mapping with "mira --project=0800 --job=mapping,genome,normal,solexa --fastq -SB:bft=gbf >&0800_Illumina_log_assembly.txt" I get 1 contig of length 4,798,671 which is pretty much 100% coverage of the genome! I've used bwa and bowtie as well which give me a more realistic ~68-69% coverage. Can anyone tell me why/how MIRA is doing this? I've played around with the setting and read the documentation but I'm at a loss! Thanks, Kasycas

09-10-2010, 04:45 AM	#2
maubp Peter (Biopython etc) Location: Dundee, Scotland, UK Join Date: Jul 2009 Posts: 1,543	I'd ask on the MIRA mailing list, but Bastien also posts here. My guess is different rules/settings for mismatch thresholds, but also MIRA isn't really doing a mapping but a guided assembly (it will extend the reference sequence(s) if it can). Last edited by maubp; 09-10-2010 at 04:47 AM. Reason: typo