Biomedical Engineering Reference
In-Depth Information
crucial to maintaining the fidelity of the databases. "Long accurate" PCR tech-
niques (62,63), using novel thermostable proofreading polymerase enzymes
such as
Pfu
, are currently capable of amplifying loci of up to ~40 kb. While
powerful, the database design should not be limited to this length by the method
of database replication, and it may be easier to enable PCR to produce ampli-
cons of somewhat longer length. One simply needs to enhance by a moderate
multiple the (amplicon) length that can be reliably amplified.
Optimized choice of amplicons may be achieved by exploiting two princi-
ples: experimental design (9,10,21) and combinatorial chemistry (44,45). Con-
tinuous variables that affect PCR reactions include the temperature of the
initiation (hot start), annealing, extension, and dissociation steps, and the con-
centration of buffer components, additives, nucleotides, primers, and template.
These variables compose a multidimensional space. A pervasive challenge in
science and technology is identifying specific values for each parameter affect-
ing multivariable processes that result in globally optimum performance and
avoid local maxima. Commercial software enables the design of experiments
that much more reliably and quickly lead to the global optimum. Noncontinuous
variables that affect PCR reactions include the identity of the template, primers,
and polymerase. An optimum combination of these molecules can be found only
by systematic screening for each. For tractable numbers of combinations, all can
be examined explicitly. When the diversity space expands beyond that domain,
"indexing" techniques are available that permit optimum performers to be identi-
fied even when in a mixture with lower performers (46). A selection of variable-
length primers can be examined, including those incorporating modified bases
(deazapurines, 2'-OMe RNA) that suppress primer consumption by dimerization.
A selection of commercial polymerase enzyme systems can be examined, in-
cluding MasterAmp
™
Taq, ThermalAce,
™
Advantage-Tth,
™
AdvanTaq
™
, and
KlenTaq
™
/Pfu. A selection of templates should be examined, including whole
viral genomes, bacterial artificial chromosomes (BACs), yeast artificial chromo-
somes (YACs), and the smallest yeast chromosome (225 kb). Analysis of the
products of these reactions is challenging due to shearing of large DNA mole-
cules by conventional sieving matrices. Pulsed-field gel electrophoresis (12,41)
can therefore be used with amplicons of this size.
3.8. Associative Search in Biomolecular Databases
3.8.1. DNA-Based Associative Search
Eric Baum (7) first proposed the idea of using DNA annealing to do parallel
associative search in large databases encoded as DNA strands. The idea is very
appealing since it represents a natural way to execute a computational task in
massively parallel fashion. Moreover, the required volume scales only linearly
