short contigs

[ru_a]

Hi,

I am trying to assemble the short reads (average length of 70 bases) from Illumina, which is from digested genomic DNA. I end up with very short contigs, average size of a contig being 150 bases. Can anyone suggest me what could be the possible reason.

Thanks.

[ECO]

More info needed. Organism, genome size/content, how much data you have, library details, instrument, chemistry revision, are just a few that could affect your question.

[ru_a]

Im working on grasses and the genome size is around 1.89GB. There are 7 libraries, pair-end with Illumina GA. Each library contains ~18 million reads and the samples were barcoded and distributed over all the libraries.

[pmiguel]

"Digested" with what?

18M 70 nt reads comes out to 1.3 GB. If your Genome size is 1.9 GB then your coverage is only 0.7X. I don't think one expects much other than repetitive DNAs to form contigs at that low of a coverage.

What software are you using to assemble the reads?

--
Phillip

[ru_a]

Digested with Pstl....I tried using GS assembler, CLC genomics and Velveth....Is there anything which we can analyse out of short contigs other than phylogenetic study.....is it possible to find some genes?

[dsenalik]

My interpretation is that if you digest with PstI and make a library out of those fragments, circularize with Illumina adapter, and sequence, then
all seqences will effectively start at a PstI site because that end is where the adapter is ligated.
In that case, you would not expect to ever get a contig larger that the read length, because you are always starting at a PstI site, never further in. The exception would be in repeat sequence regions, where contigs might overlap.

So the expected result should be short contigs, theoretically exactly equal to the Illumina read length (minus bar codes).

I've never done this, this may be crazy talk, so I am curious if this is how it works.