Galaxy |

minipolish (version 0.1.3+galaxy0)

Long reads for polishing:

Miniasm assembly to be polished:

Rounds:

Number of full Racon polishing rounds. Default: 2

PacBio reads:

Use this flag for PacBio reads to make Minipolish use the map-pb Minimap2 preset. Default: assumes Nanopore reads and uses the map-ont preset

Skip the initial polishing round:

Appropriate if the input GFA does not have 'a' lines. Default: do the initial polishing round

Purpose

Miniasm is a great long-read assembly tool: straight-forward, effective and very fast. However, it does not include a polishing step, so its assemblies have a high error rate – they are essentially made of stitched-together pieces of long reads.

Racon is a great polishing tool that can be used to clean up assembly errors. It's also very fast and well suited for long-read data. However, it operates on FASTA files, not the GFA graphs that miniasm makes.

That's where Minipolish comes in. With a single command, it will use Racon to polish up a miniasm assembly, while keeping the assembly in graph form.

It also takes care of some of the other nuances of polishing a miniasm assembly:

Adding read depth information to contigs
Fixing sequence truncation that can occur in Racon
Adding circularising links to circular contigs if not already present (so they display better in Bandage)
'Rotating' circular contigs between polishing rounds to ensure clean circularisation

CIGARs

It is important to note here something that Minipolish does not do: change/fix the CIGAR strings indicating contig overlap. While circular contigs will be connected with an overlap-free link (i.e. a CIGAR of 0M), links between linear contigs will have overlap.

So take CIGAR overlaps between polished contigs with a grain of salt. They will still indicate the approximate amount of overlap, not the exact amount.