Re:PCR? With what primers? (Score 4, Informative)

They would use random primers. A DNA hexamer (six-base sequence) is sufficiently long to serve as a PCR primer, but short enough that it would take only 4096 different types of molecule to comprise all possible sequences. Of course, we don't want the sample DNA to be plastered in our primers, so we'll pare those 4096 down to a handful, at least one of which, in any sample sequence of significant length, will nonetheless find somewhere to anneal. Once we've gone through enough cycles, it's likely that we'll have amplified at least some segment of the sample DNA. Then, getting the reaction contents purified and sequenced is simply a matter of applied microfluidics.