scaffoldig/re-assembly of mira output using quality values

[Gorbenzer]

Hi guys,
i've been assembling NGS data (454 in the past and IonTorrent now, unpaired data) with MIRA for quite some time and i usually re-assemble the contigs generated by MIRA with Lasergene SeqMan 7.

This provides really good results by finding overlaps between contigs but the problem is that this software can't load quality data from the *.fasta.qual file given by MIRA and just load the unpadded.fasta output even if it can manage other formats with quality values, like ab1 for example.

Do you have any suggestion about a way to import quality values in SeqMan or another software that will do the work??



Thank you so much for your help

[maubp]

If SeqMan can take FASTQ, combine MIRA's FASTA + QUAL files into FASTQ files.

[Gorbenzer]

SeqMan 7 can't load fastq files

[Gorbenzer]

anyone have any suggestion?

[JackieBadger]

Does it really matter if Qual is not included when you join contigs?
I would assume that you would include quality thresholds for contig assembly/Q.clipping during your Mira assembly?
The end result should be high quality contigs....Mira is pretty good at producing these (at least I know for 454), which then you can align.

Take a look at SeqManPro, or ask the developers. I would be surprised if this program can not utilize the quality values embedded within your assembly ace file to join consensus sequences.

[Gorbenzer]

Does it really matter if Qual is not included when you join contigs?
I would assume that you would include quality thresholds for contig assembly/Q.clipping during your Mira assembly?
The end result should be high quality contigs....Mira is pretty good at producing these (at least I know for 454), which then you can align.

Take a look at SeqManPro, or ask the developers. I would be surprised if this program can not utilize the quality values embedded within your assembly ace file to join consensus sequences.

Contigs generated by mira happens to have low quality ends that results in overlaps with SNPs where contigs join... at the moment i'm correcting this problem by checking everyone of them in consed and manually trimming the low quality contig end... obiously, since seqman support trimming based on quality, if it would accept .fasta.qual or fastq files the assembly would improve a lot (immagine that an overlap isn't recognized because of a long low quality end of a contig....)

[sklages]

Hard to believe that such commercial software is not capable of loading sequence formats with quality values (except for "read data").

hmm, by reading http://www.dnastar.com/t-fileformats.aspx I suspect you are right :-)
Have you asked the lasergene support about that?

Maybe you should look for another solution .. you could give gap4/gap5 from the Staden package a try. This works smooth with MIRA: MIRA->CAF->GAP.

my 2p, Sven

[JackieBadger]

You don't need to be manually trimming low quality ends.
Trimmomatic will trim your ends based on either absolute quality value of each base, or the mean across a designated sliding window.