Biomedical Engineering Reference
In-Depth Information
3.3.1. Fragmentation of the input DNA
The creation of a Biomolecular Database must involve some degree of
fragmentation of genomic DNA. While this may at first seem a very simple step,
it is in fact critical to later processing that this fragmentation step be done in a
highly controllable way. We describe several methods to produce DNA strands
of the desired length. The methods are required to produce a predictable distri-
bution of lengths, and to ensure that at least one of the resulting ends has a de-
fined sequence.
a.
Mechanical shearing
. This is a method that produces a certain size distri-
bution; however, it is not so useful in our context since the resulting ends have
undefined sequences.
b.
Reagent-less methods to create breaks
. Pirrung, Zhao, and Harris (48)
developed a nucleoside analogue whose backbone can be cleaved by long-
wavelength UV light, and specific photocleavable T analogues could be used
(analogous to the dUTP method). However, again it is not so useful in our con-
text since the resulting ends have undefined sequences.
c.
Controlled digestion of high-MW DNA by DNAse I
. This is another
method that can be used to produce DNA of a specific size range. It relies on
careful monitoring of reaction progress and does not produce specific sequences
at the ends of the fragments to enable ligation or PCR processes.
d.
Digestion of DNA with restriction endonucleases
. This offers the advan-
tage that known sticky ends are generated.
e.
"Rare Cutting" endonucleases
. These can be used to produce DNA frag-
ments of larger size. The recognition sequence of such enzymes is as large as 8
bp, meaning that, on average, DNA is cut to 1/(0.25
8
) or 65 kb. In many situa-
tions fragments larger than 65 kb may be desired; for example, complicated loci
with many introns might comprise as much as 100 kb, and that is just for one
gene.
f.
PCR methods for fragmentation
. One attractive alternative is to use PCR.
Random-primed PCR has been used to amplify the whole genome of a single
sperm (4,35,67). The challenge in using this strategy is to create long amplicons.
In principle, amplicon size in random-primed PCR is a function only of the av-
erage distance between two inward-facing hybridized primers, which is then a
function only of the primer concentration and temperature. Modest flexibility
exists in the hybridization temperature in PCR, so a fruitful strategy to make
long amplicons is to lower the primer concentration. In order to efficiently am-
plify with a low primer concentration, the primers should have a high melting
temperature (
T
m
). Increasing length and G/C content increase primer
T
m
. Ran-
dom-primed PCR can therefore be examined with novel conditions and primer
designs to maximize the amplicon length. Lengthening the random primers by
oligonucleotide synthesis is straightforward. Making the primers G/C-rich is
challenging, as G/C-rich templates are known to be more difficult to amplify
