What is the "referenceFilename" that Quiver needs to polish a assembly ?

Question

[Rox]

Hi everybody,

It's my first time using Quiver, everything is installed and now I'm trying to running it.

I have a diploid assembly, pb_268_p_ctg.fa, and I want to polish it using Quiver. So I tried something like this :

quiver --diploid /media/loutre/DATAPART1/Documents/suzukii_assembly/falcon_assembly/wolf/pb_268_p_ctg.fa

but it seem that I missed something because he keep asking for a referenceFilename

I'm not sure about that, but I use to think that the polishing step uses the original reads (PacBio reads that are longer enough to correct themselves) in order to correct the little mistakes in the assembly.

So should I use as a reference the filtered_subreads.fasta (90 Giga almost) that contains all my raw reads ? Or I'm wrong for the whole process ?

Thanks for your advices !

Cheers,

Roxane

Answer 1

[colindaven]

You need to align your reads to your assembly first using pbalign.

I have had success with this kind of shell script:

dir=myAssemblyDir

  samtools faidx $dir/$dir.contigs.fasta

  find . | grep bax.h5 > baxQuiverInput1.fofn

  source /MYLOCALPATH/smrt/install/smrtanalysis_2.3.0.140936/etc/setup.sh

  pbalign --nproc 60 --forQuiver baxQuiverInput1.fofn $dir/$dir.contigs.fasta $dir.cmp.h5

  quiver -j 60 $dir.cmp.h5 -r $dir/$dir.contigs.fasta -o $dir/$dir.contigs_quiver.fasta -o $dir/asm.quiver.gff

[lakhujanivijay]

can you please post pbalign commands for pacbio sequel data?