Optimal experiment design
The goal for this part of discussion is how to obtain as much information
as possible from a single polony slide. One obvious way to achieve this
goal is to improve polony density, which is equivalent to reducing
polony size. Polony size depends on the concentration of polyacrylamide
gel, as well as the size of PCR amplicons. We intially started with 3kb
amplicons in 6% gel, and were able to obtain >5,000 polonies with
the size of 20~50 microns in one slide. Big amplicons cover
more SNPs than smaller amplicons do, so less amplification primers need
to be used for a given set of SNPs. This provides a cost advantage
because the acrydite-modified primers account for a large percentage of
cost in polony haplotyping experiments. However, the plan with big
amplicons failed to work because of two technical reasons: (1) SBE
gives extremely weak signals when the SNP is >800bp away from the
anchor primer. (2) The amplification efficiency on large amplicons is
very low, so that even one can get many polonies within a slide, the
number of overlapping polonies is very small. In fact, only those
overlapping polonies are informative. We then switched to small
amplicons. We did experiments on CD36 gene and NOS3 gene with two set
of parameters. For the CD36 gene, the average amplicon size is 840-bp,
and the amplification were done in 10% gel. We had on average 1,397
polonies on each SNP, and 275 two-locus haplotype reads on each pair of
SNPs. We used larger amplicons (1,045-bp) but lower gel concentration
(7%) in the NOS3 gene. As expected, the polony density is slightly
higher (1,812 polonies/slide), but the average number of two-locus
haplotype reads is three-fold lower (97/slide) because of the
difference in amplicon size. These results suggest that an
average amplicon size of 800~850-bp probably provides the best
efficiency.