IEEE Spectrum April, 2015 - 16

a GooGle For dna
companies are vying to build the best
genome search engine

In 2005, next-generation
sequencing began to change
the field of genetics research. Obtaining a person's entire genome became
fast and relatively cheap. Databases of
genetic information were growing by
the terabyte, and doctors and researchers were in desperate need of a way to
efficiently sift through the information
for the cause of a particular disorder or
for clues to how patients might respond
to treatment.
Companies have sprung up over
the past five years that are vying
to produce the first DNA search
engine. All of them have different
tactics-some even have their own
proprietary databases of genetic
information-but most are working
to link enough genetic databases so
that users can quickly identify a huge
variety of mutations. Most companies also craft search algorithms to
supplement the genetic information
with relevant biomedical literature.
But as in the days of the early Web, before
Google reigned supreme, no one company has yet emerged as the clear winner.
Making a functional search engine
is a classic big-data problem, says
Michael Gonzalez, the vice president
of bioinformatics at one such company,
ViaGenetics, which was expected to
relaunch its platform in March. Before
doctors or researchers can use the data,
genomic data must be organized so that
humans can read and search it. The first
step toward that is to put it in a standard
form called the variant call format, or
VCF. As raw data, a person's complete
sequenced genome would take up about
100 gigabytes, so a database that adds
the genomes of even 10 patients per day
would quickly get out of hand. But VCF
files are more compact, requiring only
16

aPR 2015

nORTh aMERICan

a few hundred megabytes per genome,
which helps researchers find the specific
variants they want to search in a fraction
of the time. Unlike a fully sequenced
genome, VCF files point only to where
a person's genetic data deviates from
the standard, the genome originally
compiled by the Human Genome Project in 2001.
With VCF, sifting the genomes themselves for pinpoint mutations isn't the
challenge for search engine companies.

Most of these companies are allocating
their resources toward efforts to seamlessly compile supplementary information about a specific mutation from other
databases across the Web, such as the biomedical research archive PubMed or various troves of electronic medical records.
Many of these tools have finely tuned
algorithms that prioritize the results by
credibility or relevance. "You want to
be able to pull together the information
known about a mutation in that position [of the genome] and quickly make
an assessment," says David Mittelman,
the chief scientific officer for Tute Genomics, based in Provo, Utah, another company designing a genetic-search engine.
In an effort to expand the information that can be attached to a genome
under examination, ViaGenetics, based

SPECTRUM.IEEE.ORG

in Miami Beach, Fla., is making its newly
updated platform useful for researchers
who want to collaborate across institutions. With ViaGenetics' tools, researchers "can make their data available to
other users, so other people can come
across these projects, request access,
and form a collaboration," Gonzalez
says. "It helps people connect the dots
between different researchers and
institutions." This is especially helpful for smaller labs that may not have
very extensive genome databases or for
researchers from different universities
working to decode the same mutation.
Although the genomic-search industry
is now focused on serving scientists, that
might not always be the case. Mittelman
envisions that Tute Genomics could eventually serve consumers directly. People
are already demanding information
about their genomes just to understand themselves better, Mittelman
says, but most companies don't yet
consider the average person to be
their primary customer. In order to
make that shift, the tool will have
to be even more intuitive and userfriendly. "Fire-hosing someone with
data that's not easy to interpret, or
using terminology that's not standardized, has the potential to confuse people," he says. Privacy is also a major
concern for the average user; the information that Tute users upload isn't stored
permanently, Mittelman says, but users
will need extra reassurance if the platform becomes available to the lay public.
And a further evolution of the industry is in the offing. Both ViaGenetics and
Tute are hoping to be able to run the
entire process in-house-from the initial DNA sequencing to the presentation of final searchable results to users.
"The market for analyzing and interpreting genomic data is very fragmented,
like the computer industry in the 1990s,
where you had to go to separate providers to buy a video card or a motherboard and then try to put it together,"
Mittelman says. "Soon this field will consolidate, as the computer industry did."
-a lex a ndr a ossol a

Andrew BrookeS/corBiS

nEwS

http://SPECTRUM.IEEE.ORG

Table of Contents for the Digital Edition of IEEE Spectrum April, 2015