Biomedical Engineering Reference
In-Depth Information
5.2. The Key Advantages of Biomolecular Databases
The key advantages of Biomolecular Databases appear to be:
a.
Bypassing of conventional impasses
: In particular, the avoidance of se-
quencing for conversion from DNA (genomic DNA and transcribed RNA) to
digital media.
b.
Ultra-compact storage media
: The extreme compactness and portability
of the storage media—a pedabit of information can be stored (with tenfold re-
dundancy) in less than a few milligrams of dehydrated DNA, or when hydrated
may be stored in a few milliliters of solution. A Biomolecular Database is capa-
ble of containing the DNA of a million individuals (6 pedabits of information) in
a volume the size of a conventional test tube.
c.
Massive molecular parallelism
: Although one query may require a num-
ber of minutes, it is operated on vast numbers of data items (DNA strands), im-
plying a processing power of vast molecular parallelism with at least a few
hundred teraflops. The operations can operate in parallel on an entire population
of DNA.
d.
Scalability
: The technology requires volume that scales linearly with the
size of the database, and a query time that remains nearly constant up to ex-
tremely large database sizes.
e.
Limitations
: The Biomolecular Database technology is limited to applica-
tions of a biological nature (where the data are DNA or easily convertible to
DNA), and the operations are limited to logical queries in the Biomolecular Da-
tabase, associative searches, and some essential database operations. It is not
intended that the technology compete in any direct way with conventional high-
performance computers. Instead, the objective is to bypass conventional bioin-
formatics methodology by processing biological material (genomic DNA and
transcribed RNA) in "wet" media, rather than digital media.
5.3. Scalability of Biomolecular Databases Systems
The key parameters of Biomolecular Database are: (a)
N
= the number of
distinct elements of the Biomolecular Database, (b)
v
= the number of variables
(each ranging over 10 possible values) used in queries, and (c)
k
= the number of
individuals in application studies.
For our practical genomic applications of Biomolecular Databases to be
fully realized in practice: (i) the database size
N
should grow to extremely large
values (with a long-term goal of approximately 10
15
), but (ii) for these applica-
tions the number of variables
v
needs only to grow to moderately small constant
values (with a long-term goal of approximately
v
= 14), since for the genomic
applications considered only a limited number of values need to be recorded in
the information tag per database element. The relative difficulty of obtaining
