Improving genome assemblies with mummer/nucmer?

Question

[zhousun21]

Hi all,

I am trying to improve a genome assembly by aligning to reference. I am using mummer/nucmer, and it runs correctly and produces a correct out.delta file of coordinates.

Does anyone have any suggestions for how to use this out.delta file to actually make the joins, with appropriate gaps lengths and overlaps as necessary? Or, if you use a different program successfully, please let me know which one you use.

Thanks in advance for any advice.

Answer 1

[Mensur Dlakic]

It would help to know what kind of assembly you have.

What exactly do you mean by improving the assembly? Ordering the contigs and putting a certain number of Ns to connect them? In most cases that doesn't do much for assembly quality other than in a cosmetic sense, and some programs will already scaffold the genome for you if there is enough evidence.

Assuming this is a prokaryotic genome and let's say 95% complete based on marker evidence, it really won't make any practical difference whether you have 50 contigs, or you assemble them into 5-10 contigs with some Ns in between.

[zhousun21]

Thanks Mensur.

This is a eukaryotic genome which was produced by SPAdes and fairly fragmented. And yes, the goal is to order/orient the contigs and put the correct-ish number of Ns between to represent gaps. Possibly cosmetic, but still useful.

Any advice on using nucmer output or other programs for this?

[Mensur Dlakic]

Unless you have something small like yeast - it helps to provide as much information as possible - I don't think SPAdes is the best assembler for eukaryotic genomes. It may be worth trying to get a better assembly before scaffolding.

But if you really want to move on to scaffolding:

https://github.com/malonge/RagTag

[zhousun21]

Thanks Mensur. I'll look into other assemblers and RagTag.