Scaffolding with pacbio reads

Question

[guillaume.rbt]

Hi,

Sorry for this naive question but I'm a beginner in genome assembly.

I'm currently doing Pacbio reads assembly thanks to HGAP, and in result I get a set of contigs.

I was wondering if I could, and how, get scaffolds from those contigs. Is it achievable without paired-end reads?

Thanks

Answer 1

[shwethacm]

Hi guillaume.rbt,

If you are looking to reuse your PacBio reads for scaffolding, that is pretty much a dead end.

If you are asking your question because you do not want to sequence a separate PE library, then you could look to see if there short-read based assemblies and PE libraries for the same species that are publicly available and use those read data to scaffold your contigs. You could get lucky sometimes by just looking :)

If you are strictly looking for non-paired-end based scaffolding, there are orthogonal approaches like BioNano and Hi-C based (super)scaffolding. BioNano is based on optical mapping using restriction sites, and Hi-C is based on spatial organization of chromosomes.

[guillaume.rbt]

Hi Shwethacm,

Thank you very much for your thorough answer. I understand it is more difficult than I thought to get scaffolds.

Actually I've just found out that I have a PE library for some strains (I work on an haploid fungi with a small genome, around 40Mb)

How do you think I should use those data? Can I use it on my assembled contigs to get scaffolds, if so do you know a tool for doing that?

Or should I redo an assembly based on an hybrid approach, taking both long and short reads at the start? And do you know which assembler would be the best in your opinion for this task?

Answer 2

[shwethacm]

If you already have contigs, then you can put them into scaffolds using many tools. Re-assembly along with PE data is not necessary - you will not be gaining anything! Here's a paper that outlines many PE scaffolding tools: https://genomebiology.biomedcentral.com/articles/10.1186/gb-2014-15-3-r42 (it could be outdated though! )

[guillaume.rbt]

Thank you for the publication, I will test some of these tools :)