de novo assembling using NEBNext™ dsDNA Fragmentase™

[hi-koike]

Hello all,

I'd like to know if anyone has tried de novo assembling from a library made by gDNA fragmentation NEBNext™ dsDNA Fragmentase™.

What software(s) are good for the de novo assembling of the library reads?

I have got the data of Illumina Miseq 2x250bp from the library made by the fragmentase. I found that the fragmented dsDNA ranged from 200 bp to 1000 bp by observed using gel-electrophoresis. There seemed no peak in the range.

I'd appreciate your idea.

[luc]

Hi,

are you not using any type of enrichment for specific fragment sizes - gel extraction or double-cut bead cleanup (e.g. with Ampure breads?).

[hi-koike]

Hi Luc,

No we did not. Because the amount of the obtained DNA were low and just enough for the recommended amount for Library construction, we decided not to do any enrichment for specific fragment sizes prior to the construction. By the analysis using bioanalyzer, we could not find any peak.

I guess during the Library construction, smaller fragments were enriched comparing to longer fragments, but we do not enriched the fragments of specific sizes.

I performed FLASH without any restriction on sizes, I could obtained the combined reads larger than 251bp. However I could obtain the combined reads for only 10%.

[winsettz]

Can you at least tell us if the species you are assembling has a complex genome or not? For instance, intergenic repeats, chromosomes and the like?

[hi-koike]

Hi winsettz,

We are assembling a fungus which might have a genome size around 20 Mbp and
having more than 25 chromosomes.
We have a reference genome of a different strain which has been assembled using 454 data.

I would try to map the paired end data onto the reference genome.

[luc]

Hi hi-koike,

if there is another chance, I would use a sonication protocol next time for fragmentation. Enzymatic reactions like fragmentase are inherently very much dependent on DNA quality and correct quantification. Sonications tends to be a lot more reproducible - at least on the Covaris. If there is another chance, I would also do a size selection ad adjust the library prep protocol to the lower input amounts (likely this only means using less adapters perhaps less ligase).