Assembly miniasm - aechchiki/SIB_LongReadsWorkshop_Zurich17 GitHub Wiki

[Miniasm](add links) is not a real OLC assembler. Is is actually performing only the Layout (L) step. Consensus (C) step is just not computed and Overlap (O) of reads is expected to be provided on input.

The assembly does not require correction of reads either, which is going to save a lot of time as well, but it will cause a big nucleotide level errors of the resulting assembly.

Wo compute Overlaps between reads using program minimap2, currently the fastest solution. minimap2 is a mapper, it's able to map read to reference, therefore when we map reads to reads, we need to specify the reads twice. The -x ava-pb are parameters recommended for computing overlaps between reads, for other options look at the [repository](add link), or [manuscript](add link).

minimap2 -x ava-pb <reads.fq.gz> <reads.fq.gz> | gzip -9 > <overlaps.paf.gz>

Now we got computed overlaps between reads (.paf file), therefore we can compute the string graph and it's Layout using miniasm

miniasm -f <reads.fq> <overlaps.paf.gz> > <contigs.gfa>

and finally convert the assembly graph (.gfa file) to classical fasta format using basic unix utilities

## convert .gfa to .fasta
awk '/^S/{print ">"$2"\n"$3}' <contigs.gfa> | fold > <contigs.fa>

❟ How many contigs there are? What is the genome size? Is it better or worse then Canu? There is no report produced by the assembler, you are alone with UNIX tool.

Consensus using Racon

Racon is a fast, stand alone module for computing consensus step of genome assembly. This module was build with OL steps expected to be computed by Minimap and Miniasm respectively. Using this three programs approaches very closely in quality the state of art assembly methods, but the computational speed in several orders of magnitude faster.

In this tutorial we do not cover Racon, however, this might be the solution for enormous plant genomes, or project where a huge amount of genomes are sequenced.

There are sudo privileges on the accounts you have, the installation should be pretty straightforward, if you really want, you can try to install Racon and run it.

Next

Learn more about Polishing.

Finish this section, go to Checkpoint

Go back to tutorial assembly using Canu

Go back to Table of content.

⚠️ **GitHub.com Fallback** ⚠️