Biomedical Engineering Reference
In-Depth Information
with the database size. However, there were further technical issues to be re-
solved. For example, the query may not be an exact match with any data in the
database, but DNA annealing affinity methods work best for exact matches. Reif
and LaBean (54) described improved biotechnology methods to do associative
search in DNA databases. These methods adapted some information processing
techniques (error-correction and VQ coding) to optimize input and output (I/O)
to and from conventional media, and to refine the associative search from partial
to exact matches.
Reif and colleagues (55) developed and then experimentally tested a
method for executing associative searches in DNA databases of encoded images,
and this method was tested using an artificially synthesized DNA database. Prior
to that project, the idea of using DNA annealing to do parallel associative search
in synthetic DNA databases had never been experimentally implemented. They
detailed a study involving the design, construction, and testing of large data-
bases for storage and retrieval of information within the nucleotide base se-
quences of artificial DNA molecules. The databases consisted of a large
collection of single-stranded DNA molecules that was immobilized on polymer
beads. Each database strand carried a particular DNA sequence, consisting of a
number of sequence words drawn from a predetermined lexicon. They made a
number of experimental databases of artificially synthesized DNA sequences
designed for encoding digital data, scaled in increasing sizes. Each DNA strand
of the database is single stranded, and encodes a number that provides the index
to the database element. They used an extensive computer search for the design
of the DNA word libraries, to ensure a significant Hamming distance between
distinct words and allow for annealing discrimination. They constructed their
largest synthetic databases in two phases. In the first phase they constructed an
initial DNA database by combinatorial, mix-and-split methods on plastic mi-
crobeads. This constituted by far the largest artificially constructed synthetic
databases of this sort. The next phase was the development of a construction
method for much larger synthetic databases by combining pairs of the synthetic
database strands so as to square the size of the database to approximately 10
15
distinct data elements (each represented redundantly by over 10 identical strands
of DNA). Even with this greater than tenfold redundancy, the DNA database
using this construction method is extremely compact, and requires only 10 mil-
ligrams of DNA.
3.8.2. Associative Search via PCR
PCR methods can be used for associative search queries in Biomolecular
Databases (in particular, on the words of the tagged portions of the Biomolecu-
lar Database strands), using known and modified PCR techniques previously
developed by Reif and coworkers (55). They describe experiments for executing
