Info 202 Fall 2011

To gain momentum 23andMe initially targeted people with Parkinson's disease**. Based on users' genetic test results they were provided with recommendations and encouraged to self-report them back to the service. 23andMe also allows users to share questions and results with each other and experts, and elicits further data from them through bite-sized surveys. All these data-collection strategies allow 23andMe to create an increasingly dense web of data correlations, and keeps its users connected as continous sources of data and funding. 23andMe stores a mix of dynamic and static data, including its users genetic data, survey data (family history, personal history, location, etc.), and self-reported treatment results, as well as internal and external scientific research. The organizational model presented to the user centers around health, ancestry, community, and research.  The health section contains information on disease risk, carrier status, drug response, and traits. 

The data is provided by users, but in the end it is 23andMe's scientific team that makes the determination what data they collect and/or leverage from users. They also conduct the statistical analyses to identify patterns resulting in new scientific data about the source data. 23andMe does not appear to share its methodology or raw data, to protect their user's privacy and to protect its trade secrets, but it does share the status of its research, and publishes its studies. Information associated with the entries presented to the user are continuously analyzed and organized as new information becomes available. This affects the organizational structure presented to the user, since new conditions can be added, or disease risks can be modified based on new research. A good example is Parkinson's disease, which is in an early research phase and still largely driven by crowdsourced data from the '23andWe'  initiative. Therefore the data is still incomplete and unstable and the associated predictions more likely to change.

CureTogether is another promising service that was heavily inspired by the QS movement. It crowdsources drug evaluations by its users to make predictions on their efficacy on hundreds of health conditions. The service allows users to define the conditions that they want to track and also allows them to contribute primers, guide, and videos. As a result the service is highly relevant and engaging, and is designed to evolve with its users.  CureTogether has partnered with several research universities and UC Davis is comparing CureTogether's data against published research, to establish how representative its data is, with the aim of establishing it as a credible scientific source. 

Interestingly, a users' 23andMe data can be exported and reimported into CureTogether, enabling users to leverage their personal data in different systems which analyze their data against their collection of data.

** Another factor, not mentioned by the article, may have been the personal interested of Sergey Brin, the co-founders's husband and an investor in 23andMe, in Parkinson's research.