454 Rapid kit in metgenomics

[v_kisand]

Hi all,

We use Roche Rapid kit for environmental metagenomics. People working in this area probably have experienced that total community DNA is usually not with as highly good quality as compared to genomic DNA from cultures, tissues etc, some DNA is fragmented after purification. So we have some degradation of our DNA always. What we have observed is that we get lower quality of reads when proportion of degraded DNA is higher, and not so much related to yield of DNA. Well quite obvious, right?

My questions can we expect to worsen or get better results when we use large amount of DNA in Rapid kit? Could we get even worse results by some preferential sequencing of shorter fragments etc.?

v.

[zhengz]

Hi kisand,

I feel it unlikely to be related with the amount of DNA. Rather, as you already indicated, the preferential amplification of shorter fragments during emulsion PCR.

The amount of PCR products generating from shorter fragments will be much higher than longer ones after 50 cycles!! This will end up very strong light signal during sequencing, which eat up the nucleitide flows and may also cause problems in signal normalization during signalProcessing.

Just a thougth - maybe you can try to size-select your library twice.

/ZZ

[v_kisand]

Meanwhile we tried to increase amount of DNA for Rapid Kit and so far results improved!
Only explanation I have, DNA quantification is always a bit problematic issue...

[zhengz]

Good that your results have been improved.

I have no experience with the Roche Rapid Kit. I only use home-made Y adaptors and qPCR quantification. In my hand, the amount of DNA is not a problem - as I actually start with low biomass metagenomic samples where the amounts were no able to be quantified by Qubit. I don't do emPCR titration. The results for me have been good too, once I got 1.7 million passed filter reads with median length 430+ bp, totally 700Mb.

I think qPCR quantification would be better than FAM measurment. I frequently pool 4 to 16 samples, and yield for each sample has been very even (<5% difference) or close to expected (e.g. I allocate 20% yield for one sample and 5% for another...)

[v_kisand]

there are some more threads concerning DNA but I just write my experience here:

in most cases it is possible to get things working but I doubt that you know _exact_ concentration of DNA (e.g. Rapid Kit protocols tells you that take 500 ng of DNA)... When DNA is originating from the mixed community DNA in the range of nanograms containing some coextracted compounds, partly fragmented etc. each method may result id different conc, sometimes really different. With very defined stuff, e.g. plasmid, pure isolate chomosomes and with a bit higher concentrations it works. Anyway as in most cases excess of DNA does not harm down stream applications I go for most conservative estimation, quite often this is bioanalyser or simple agarose gel quantification... Fluorometric bulk measurements tend to give higher values.

qPCR quantification is a good idea when you have right primers...

[zhengz]

No, I don't know the concentration. As I said, they were not able to be quatified by Qubit High Senstivity, meaning that the amount were less than 1 nanogram.

I am interested in this thread because of "metagenomics", where I think sub-nanogram of starting DNAs could sometimes be the cases and qPCR quantification is the way to go.

I have one sole QC step for the entire procudre - qPCR quantification to see how many molecules are effective for emPCR and agarose gel analysis using qPCR products to check the size distribution of library and to check if there exists adaptor dimer. Yes, one might wonder if they fail the QC, should I repeat the library construction? I would repeat AMPure selection using less volume of beads to get rid of adaptor dimer (this happned when I used two adaptors, and never so far when I use Y adaptor). In case of I need to start from the post-nebulization step (never happened so far), instead of using the Rapid Kit, I bought some regeants that could used for preparing hundreads of libraries at the cost of <1000 $.