No announcement yet.

Personalized Genomics Challenges: Trust Google with my Genome?

  • Filter
  • Time
  • Show
Clear All
new posts

  • Personalized Genomics Challenges: Trust Google with my Genome?

    Personalized genomics companies Navigenics, 23andme, and Knome have all been announced in the past year or so, promising to offer customers a peek into their genomes. The goal of these companies is simple. Take advantage of ultra-high throughput SNP genotyping and sequencing technologies to bring the consumer “personalized genomics” (PG) information. This could take the form of hundreds of thousands of SNP genotypes, or even whole genome sequencing. Here I will touch on a variety of issues facing personal genomics startup companies, and perhaps start some discussion here on SEQanswers.

    Figure: The Role of a Personalized Genomics Provider (full PDF version in post attachments)

    The foremost concern of most people when approached about having their genome sequenced is privacy. The non-scientists in my circle of family and friends are remarkably aware of the privacy issues brought about by this new type of health information; the horror scenario of genetic discrimination by insurance companies almost always comes up. The impact of exposed genomic data is limited only by a sinister imagination, one could imagine ordering synthetic amplicons that correspond to someone’s CODIS profile and then framing them for a crime. Privacy will be a major struggle for PG companies as they will almost certainly utilize the internet for communication with customers.

    I have privacy worries that segment into two different categories. The first of which is basic physical security of data. Call me paranoid, but I would not allow my genome to be web-accessible with the only requirement being a plaintext password. Some higher level of security will undoubtedly have to be implemented, such as a two factor authorization system like the SecurID tag commonly used for corporate networks. I would personally prefer some sort of biometric measure (fingerprint, etc), however I doubt the feasibility of this. The safest form of access that I can come up with would be a completely standalone bootable operating system which is designed only to launch a suitable genome browsing utility. With the amount of spyware found on Windows machines, offline access on a compromised personal machine could be just as dangerous as network access.

    The second category of privacy worry relates to the title of this article. While I will save a detailed discussion of potential PG business plans for another article, there is a very real danger that one’s data is insecure even when it’s sitting “securely” on such a company’s server. As these PG companies are all for-profit entities, it’s very worthwhile to question the continued privacy of genomic data when the company is faced with lucrative financial incentives to share that data with partners. Think of your favorite web-based email account (I typically use Gmail), and notice how all your communications are parsed and relevant ads are shown to you alongside your free email account. Now imagine that you’re CEO of a personal genomics startup looking to generate revenue beyond a $2500 one-time SNP chip that is probably less than 25% margin. The real payday will come from pairing medical service providers with ultra-targeted customers. A customer base like this is a drug company’s dream; the ability to place ads for their latest wonder drug directly to those potential patients which have a favorable genomic profile. My wife pays up to $1 for a customer click from a targeted Google ad for her jewelry business, imagine what a drug company would pay to a PG company for a “partnership”. I’m not suggesting this is a horrible thing, but I would want to ensure my customers that there is no potential for identity recognition or abuse by advertising clients. I would argue that even sharing an innocuous number like “ad impressions” with a drug company would be a dangerous road to walk down. With 23andme co-founder Anne Wojcicki being married to Google co-founder Sergey Brin, this scenario has undoubtedly been paraded to investors.

    Data ownership
    A recent article in the Wall Street Journal describes the launch of Navigenics’ initial SNP genotyping service. In this article, and a subsequent analysis by David Hamiton at Venture Beat, Navigenics has tipped their cards at a “lock-in” business model. Their service will charge the customer $2500 for an initial test (an Affymetrix SNP chip), and offer an interpretation about the small variety of diseases for which semi-concrete association data exists. The lock-in comes in the form of a subscription, for which the interpretation will be kept current as advances in genomic understanding are made. A perfectly understandable business plan, however it is unclear the breadth of data access non-subscribing customers would get. Hamilton at Venture Beat suggests that raw data will be available, however some waffling on Navigenics’ part seems that they ultimately want to be the curators of your genome for a price. Alternatively, Knome states outright that they will give you immediate full access to your data, enabling you to take it to any analysis provider.

    I feel that full access is essential, and thus question the business plan that includes holding customer data hostage for a fee. I will predict here that there will be an open source repository of genome-disease correlation that accepts personal genome data. Sort of a UCSC Browser + OMIM + dbSNP + Google Maps on steroids for the non-scientist. If Wikipedia can survive on donations, such an effort will undoubtedly come out of the biological community, perhaps even funded by NIH.

    Where does the money come from?
    With Illumina advertising >500k SNP chips for <$300, high-throughput SNP genotyping is virtually a commodity. It’s only a matter of time before one of the many genotyping service laboratories realizes that there is a market for personalized genomics and starts offering testing directly, rather than subcontracted through 23andme or Navigenics. This will drive consumer testing costs down, leaving behind the difficult task of distilling the information and presenting it to the consumer in an understandable manner. Although it is unclear who the payers would be for such services, there are undoubtedly a variety of extremely valuable services a PG company could provide such as:
    • Selling detailed reports designed to provide a customer’s physician with immediately relevant genomic data, referenced with appropriate medical studies and potential treatments.
    • A reference for genetic counseling services. Parents-to-be would almost certainly want to discuss joint risks to their unborn children.
    • Disease risk reports, up to and including an overall “genetic health” metric. Cancer susceptibility, diabetes risk, etc.
    • Interface with the researchers, database access. Selling access to an anonymous database for research purposes.
    This type of health information is more forward looking than any of us have ever experienced, which could lead to customer anxiety regarding the interpretation. The true meaning of genomic data must be explicitly conveyed to patients to prevent undue panic.

    The Future
    It is an exciting time in genomics, when data generation technology has outrun our biological understanding. The personal genomics startups discussed here will be pioneers in the areas of genetic privacy and customized healthcare.

    Attached Files

  • #2
    Personal genomics era and issues

    Below article describe privacy issue especially with health insurance company, and potential economic impact.

    Maybe U.S. will have a public health care system in near future .


    • #3
      Ad-genes the gene targeted ad bidding system. I think that is patentable.