How Can I Increase The Length Of Existing Genome Contigs Using Paired-End Reads?

Question

[aswad87]

Hypothetical scenario:

I am interested in a genomic region ~100 Kbp long. I have Identified (via BLAST of a similar sequence) approximately half of this in the form of contigs from a draft assembly. some of these contigs overlap, but they are mostly discontiguous with respect to the whole sequence.

Does anyone know of a way to use the paired-end reads to 'fill-in the gaps', using the known contigs as a starting point? The idea would be to avoid reassembling the whole genome. The only reference I have is from a very distantly related sequence, so cannot be relied upon for mapping.

Thank you in advance!

Answer 1

[Istvan Albert]

This sounds like a basic scaffolding approach. Map the reads using the contigs as your target and see how many read pairs map to different contigs. Those will indicate contigs that are ordered in a certain way. This will work if the contigs are no more than an insert size away from one another.

Probably there is a tool that can do this for you. Now that I wrote that you could alsor perhaps attempt to assemble you paired end reads, then once done assemble again the old contigs with the new ones.

Answer 2

[Michael R. Crusoe]

We use ATLAS_GapFill for this task (as does the Broad, and others)

Answer 3

[Leszek]

I have used SSPACE with success. It accepts multiple libraries (pair-end, mate-pair) and gave me quite good results. It also fills the gaps. For gap closing you can also try GapCloser from SOAPdenovo package.

Answer 4

[aswad87]

Thanks for your answers! I will try all the suggestions!