How To Use Unmapped Reads To Reconstruct Strains Specific Genome Segments?

Question

[William]

Are there any common approaches to make use of unmapped reads to reconstruct strain specific genome segments?

Are these just your normal denovo aproaches do you somehow also take the mapped reads and the reference in to account? Maybe for example read pairs in which one reads maps and not the other? Or reference assisted assembly?

And is this at all possible with Solid (fragment and PE ) sequencing data? A denovo assembly or reference assisted assembly?

[Fabio Marroni]

Both your approaches are reasonable. If you have a relatively small genome, enough CPUs and RAM, and good libraries, it's easier to perform a complete denovo assembly and then identify additional segments in your strain.

Answer 1

[Philipp Bayer]

There was a recent paper which did this with a ton of rice accessions:

Exploring the rice dispensable genome using a metagenome-like assembly strategy

We proposed an assembly strategy incorporating SOAPdenovo [21], Meta-IDBA [22], and GICL [23] to assemble the dispensable sequences (“Materials and methods”; Fig. 1a). Both reads of a read pair of which at least one read could not be mapped to the reference genome were extracted to do de novo assembly (“Materials and methods”). In total, 395,891,829 read pairs (7.0 % of all the read pairs) were collected, of which 628,163,124 reads (79.3 %) could not be mapped to the Nipponbare genome

Answer 2

[fridhackery]

I don't know if its common, yet, but certianly doable: A: Collect read pairs where at least one one read is mapped