Binge-watching “House of Cards” isn’t going to cure cancer, but one bioinformatics researcher says there is plenty to learn from the algorithms Netflix uses to personalize your queue.
The same “matrix factorization” algorithms deployed by the streaming media provider to narrow TV and movie recommendations for users can provide benefits for genomic cancer data, writes Elana Fertig, an assistant professor of oncology biostatistics and bioinformatics at Johns Hopkins University, in The Conversation.
Thanks to initiatives like the Human Genome Project and the Cancer Genome Atlas, cancer researchers now have access more to genetic data than ever before. The Precision Medicine Initiative, launched by President Barack Obama in 2015, aims to personalize medical treatment, and the Cancer Moonshot is looking specifically at ways technology can drive a more precise approach to cancer care.
But identifying the specific genes responsible for each type of cancer is not that simple, because scientists need to sift through thousands of data points to identify a single gene linked to one type of cancer.
That’s why Netflix’s personalization algorithms—which rely on matrix factorization to identify common ratings and patterns between isolated groups of users—are a convenient solution.
“The same process can work in cancer,” Fertig writes. “In this case, the measurements of gene dysregulation are analogous to movie ratings, movie genres to biological function and users to patients’ tumors. The computer searches across patient tumors to find patterns in gene dysregulation that cause the malignant biological function in each tumor.”
Fertig and her colleagues have already developed an algorithm that mirrors this approach, but that’s only one battle in the war against cancer. Going forward, cancer research needs “a new generation of scientists who can bridge mathematics and statistics" along with powerful computer programs that can predict the long-term outcomes of specific treatments.