different PCR cycle number for different microbiomes?

[petralutra]

Hi,
I'm doing 16S amplicon pyrosequencing of oral and vaginal microbiome in wild mice. Samples have low DNA concentration (between 0.0001 and 0.5 ng/ul based on qPCR with 16S fusion primers; Qubit values are ca 10-fold higher). I would like to use the smallest possible number of PCR cycles to avoid biases, contamination etc. Vaginal samples have 10-100x lower DNA concentration than oral samples. Would it be ok to use different PCR cycle number for O and V if I plan to compare them? I can get nice bands on gel with 34 cycles for O and 38 for V samples using Kapa hifi (there is a paper on house mouse microbiome where they had to use even 44 cycles).

There are also quite big differences between particular samples so I'm even thinking of dividing them into 3-4 categories (according Cp in qPCR) and amplify them with different number of cycles. However, I did not find such approach in literature, so I'm not sure if it is ok? Better approach would be to dilute samples to similar concentrations and to use the same number of PCR cycles, but considering very low concentrations of my samples I really don't want to do that .

Many thanks for your opinions!

[scott.yourstone]

You might want to check out this paper where each original DNA template gets a unique tag (similar to barcoding) prior to PCR and sequencing. After PCR and sequencing reads with the same tag can be grouped (because they come from the same template molecule). A single consensus sequence can be built from each group thereby reducing both sequencing error and PCR bias.

http://www.nature.com/nmeth/journal/v10/n10/abs/nmeth.2634.html

I know that a lab here at UNC has been using this protocol on lung microbiome samples where DNA concentration is also very low.