Genomics Needs A Killer App

Genomics is an incredible technology for newborn and pre-natal screening, investigating the cause of a rare condition, and treating cancer, but for most healthy people it’s still not particularly revolutionary.

Imagine if instead, with the information locked in your DNA, we could predict what diseases you’ll have and find effective and precise treatments, with no side-effects. What’s standing between where we are now and that vision is a better understanding of DNA and what it means in the context of health and medicine.

Genomics has a distinct 1994 feel. In 1994 the Internet was confined to people that no one would ever call mainstream. The user interfaces were terrible, your coworkers weren’t on it, and high-speed connections weren’t available to most people. But more importantly, it wasn’t clear why better fiber and more websites would be useful to the average consumer. There wasn’t that much to do online. The surplus of AOL CD’s distributed to America by mail remains one of the saddest memories of the 90’s. That’s where genomics is now.

Genomics has undergone a major shift in the last year. The drop in price in the past decade from $3 billion to $1,000 to sequence a genome is the 10X force likely to cause an inflection point in a number of health-related industries. But despite the price, right now genomics is a niche product in large part used in and promoted by academia.

Only 100K-1M full or partial genomes have been sequenced* and much of that has been driven by research investment rather than consumer demand. Applications of genomics are only just starting to be commercialized by cutting edge diagnostic companies and explored by innovative teams in biotech and pharma.

Genomics has undergone a major shift in the last year. The drop in price in the past decade from $3 billion to $1,000 to sequence a genome is the 10X force likely to cause an inflection point in a number of health-related industries.

Of course, the market for genomic tests holds great promise. It is expected to grow 10 fold next year. But an entire industry is wondering: is this a temporary reconfiguring of technologies to suit existing demand or is this is the beginning of sustained explosive growth? Will demand for genetic tests, genomic-based R&D, and molecular digital health grow into the next Internet and create several rocket-propelled unicorn companies worth billions?

To answer this question it is important to understand that the current 10X market growth, though remarkable, is still in the early adopter sad phase.Genomics, even in rapid growth mode, still only appeals to a few niche clinical categories: people that are either already sick, usually with complex conditions, or want to be screened for a small number of known disease markers.

To be the next Internet, genomics needs its “light bulb moment” – the singularity where the technology reaches the point where applications can be built and deployed to the mainstream market leveraging the infrastructure built for and by previous applications.

One App Away

Before the light bulb, electricity was also a niche product. It was so niche that no one used it and generating electricity and wiring houses made no sense. Electric wiring became necessary as a means to power Thomas Edison’s iconic invention. The light bulb was the killer app that drove infrastructure investment. Side note: Edison’s other killer app was the electric chair.

Initially only JP Morgan had electric wiring in his home – he had an on-prem generator. But demand for the light bulb skyrocketed because it was at least a 10X improvement over kerosene lamps, candles, and houses burning down. As the light solution moved from a vertically integrated, on premises generator to “cloud” electric generators powering entire neighborhoods and cities, the electric grid was built.

To be the next Internet, genomics needs its “light bulb moment” – the singularity where the technology reaches the point where applications can be built and deployed to the mainstream market leveraging the infrastructure built for and by previous applications.

The result was a network that other appliances could now plug into without the initial setup costs paid by JP Morgan and other early adopters.  An entire ecosystem of apps could now be designed and built by entrepreneurs because the distribution system and infrastructure was already in place. The initial technology, electricity, was driven by the killer app to mass distribution where network effects kicked in.

Without the driver app, the underlying platform technology and the entire app ecosystem would not have existed. The technology revolution may not have happened.

Healthcare will dramatically change when it becomes radically easier to build and deploy genomics applications. Research, development, clinical diagnostics, pharmacogenomics, cancer screening, genealogy, agriculture, nutrition, and industries and fields that aren’t even on our radar will all be impacted. But this will only happen when the genomics light bulb drives infrastructure investment.

Hurdles 

Despite the dramatic drop in price, right now even enthusiasts don’t have their whole genome sequenced. The reason is kind of a big one: We don’t know the biological meaning of most genomics markers. What we do know comes from a relatively small number of databases, collected by scientists, which are used as the reference codex to determine whether someone is healthy or sick.

Unfortunately, because the field is so young, the databases used by researchers, clinicians, and companies alike are often compiled in the world of academia and not intended or suited for diagnostic use. Reliability, usability, security, and scale of the underlying information have not been a priority. This is a huge problem.

The driver app of genomics will be an application that helps the industry “understand the genome”, in other words, use genomic information that we currently don’t know how to apply to medical problems.

For this reason many genetic tests performed today are likely to be incorrect and/or irreproducible. The labs might be using the most cutting edge technologies, but when operating in a world where the map is missing information, wrong turns are inevitable. This problem can only be solved by standardizing available data and filling in the gaps in the existing references.

Luckily, since more and more data is being generated from both sick and healthy individuals, there is an opportunity to improve the data layer upon which diagnostic applications are built, and thereby improve the accuracy of genetic tests, for those labs that learn from the information they, and others, gather. The value network to collect and disseminate data (think: Bloomberg Terminals for financial data) is in its infancy and the infrastructure necessary to support it is just now being built (and the startup where I work, SolveBio, hopes to help).

The reason the genomics data layer is a problem with global consequences is that until we understand how to more accurately interpret DNA sequence,genomics will be confined to the early adopter niche market. The driver app of genomics will be an application that helps the industry “understand the genome”, in other words, use genomic information that we currently don’t know how to apply to medical problems.

The Driver App: The Genomic Network

Genomics isn’t normally described in terms of network effects, a dynamic usually reserved for photo sharing, communication, and the sharing economy, but genomics is actually an entirely network-based field. Collecting genomic observations improves one’s understanding of the genome, which in turn improves the ability to deliver value through the application that is based on genomic information, which then enables further data collection. Applications that facilitate the collection and dissemination of this kind of encyclopedia of genomic information that can be referenced by other applications will drive the industry to new heights.

Traditionally, much of this information has been distributed through academic publications. Many companies curate papers to extract valuable information for clinical genomic and R&D applications: IngenuityBiobaseThomson Reuters, and others.

We are starting to see more and more commercialization of medical information, directly from patients, outside of academic publishing. IMS Health aggregates and monetizes drug prescription and insurance claims data; 23andme recently sold access rights to its Parkinson’s cohort database to Genentech; and Foundation Medicine just announced a deal with a pharmaceutical company to use it’s cancer genomic database for research.

Companies and hospitals developing and deploying diagnostics tests and research applications, collect and curate genomic data as part of their intellectual property.

The industry is still young and the volumes are still small, but as adoption grows, many labs are likely to develop IP advantages by having access to data that their competitors don’t. One example is Myriad Genetics that has developed, and kept private, a database of BRCA gene variants. Other testing companies, such as Invitae and Counsyl have the opposite approach of submitting the data they collect to public repositories such as ClinVar.

Since no one company or hospital is likely to observe all genomes and conditions, the genome can really only be understood in full by exchanging observations and experimental results.

One Man’s App Is Another Man’s Platform

The Internet serves as a better analogy for the potential of genomics than the electric revolution because the light bulb application was clearly built on the electric grid platform and was not itself a network, platform, or infrastructure – it had little vertical influence on other appliances.

In contrast, the Internet grew through viral network apps, like Hotmail and ICQ. Email and instant messaging drove infrastructure investment in the underlying network, but they also were networks themselves. Many email and IM applications ended up being platforms and infrastructure for future applications (email-based electronic payments, like the early PayPal, are good examples).

The exponential growth of applications that became possible thanks to the Internet changed every industry; the genomics network has that potential. Right now genomics is starting to be used for cancer drug targeting, family planning genetic screening, the diagnosis of rare diseases, and R&D. The field has been driven by a 1000X improvement in sequencing technology. If there is a similar improvement in the way the data is interpreted, the possibilities are extremely exciting. Drugs will be developed precisely for the molecular profiles of an individual’s ailment.

Diseases will have new names: they will no longer be referred to as vague groupings of symptoms but rather as exact molecular pathologies (instead of “colon cancer” it will be the exact genes and pathways disrupted). Drugs won’t have side-effects because they will be targeted, and clinical trials won’t cost literally billions of dollars because pharmaceutical and biotech companies will be able to more accurately stratify populations and predict who will benefit from the drug and who won’t.

We know entirely too little about complex human conditions like autism and fields of study like nutrition. Those fields are currently viewed through the prism of population-wide data and theories. The advent of the genomics network will spawn applications in each of these arenas that will allow doctors, and eventually consumers themselves, to treat individuals as individuals.

* An older genotyping technology, microarrays, has been employed at greater volume by labs such as 23andme. This technology does not sequence DNA but rather probes for the presence or absence of known markers.

Mark Kaganovich is the CEO and cofounder of SolveBio. Follow him on twitter at @markkaganovich.