How Google is Helping to Treat Autism
In this video, PBS reports on the progress made by the ambitious MSSNG project. The project is an unlikely collaboration between the Autism Speaks organization, the Toronto Hospital for Sick Children, and Google with the goal of mapping the genomes of ten thousand people affected by autism and storing them in a useful database. Autism Speaks provided fifty million dollars to fund the project, while the Hospital for Sick Children had the technology to map DNA from blood samples at low cost. The MSSNG database contains the DNA of thousands of people with autism as well as their siblings and parents. This is a huge amount of data that had to be efficiently analyzed in order for researchers to start identifying the specific gene mutations that correlate to autism. Working with Google, all of the data was uploaded to the cloud and could be analyzed by the powerful search engine they developed. The search engine went through trillions of rows of information to find patterns and answer some of the researcher’s questions. It has already helped them to determine the mutations that caused autism in individuals and they hope to use this information to identify the disorder earlier in children so that treatment can begin earlier. Google also hopes to expand this method to finding the causes of and improving treatments for other diseases such as cancer and more uncommon diseases that are related to genetics.
We have discussed how search engines prioritize pages for top search results using PageRank and Authority. This kind of ranked search that Google normally performs is very different from what is necessary for the MSSNG project. It would be very interesting to learn more about how the Google search engine for the genome database analyzes all of the data to find meaningful patterns. The search engine has to identify the DNA of the individuals with autism and then compare the data to the genomes of their families and others without autism to find mutations that may be the cause. The search engine has to do this thousands of times as there are approximately two thousand three hundred individuals currently in the database. And of course each individual genome by itself is a massive amount of information that has to be analyzed. The ten thousand affected individual goal is projected to be reached by next Spring, but this project has already made important discoveries about the genetics that affect autism and will continue to make progress towards better treatments.
http://www.pbs.org/newshour/bb/search-engine-help-unlock-autisms-secrets/