What's the best practice to do scaffolding?

Question

[xiaoyezao]

Hi,

Can anyone help about scaffolding?

My situation is: I have a quite good genome assembly (contig N50 = 5M) which was generated from 80x pacbio long reads. In addition, I have 100x illumina pair-end short reads.

Do I need to do scaffolding on this assembly? What's the best practice to do it?

[Carambakaracho]

Do I need to do scaffolding on this assembly?

nope. There's always a slight chance you might be able to merge one or the other scaffold but it's unlikely. You can try to polish the assembly, though.

Answer 1

[colindaven]

Paired end short reads are from small fragments, see for example here.

https://www.ecseq.com/support/ngs/what-is-mate-pair-sequencing-useful-for

If you had long range information from say good 3-10 kbp libraries (its' really hard to generate good ones) then that was useful in the pre-long read era.

With your long pacbio reads you should probably be pretty happy with 5 MB contigs. Perhaps nanopore or HiC could be added to make a slightly more contiguous genome.

Another approach might be reference based scaffolding (eg program Medusa and others), assuming you have a closely related excellent reference sequence.

[Juke34]

+1 for HiC or Nanopore.
You can use 10x too