Do you have a friend who has beaten the odds and survived cancer? I do.
When I was in dental school, my best friend’s husband was diagnosed with melanoma and given 6 months to live. Today, over twenty years later, he is still enjoying life. Imagine if there was a technology that could show us why some people respond to treatment to survive beyond their initial prognosis and others don’t?
Perhaps Ayasdi’s platform will hasten our understanding of these mysteries.
Using Math for Almost Instant Pattern Recognition
Ayasdi is applying topological data analysis (based on the mathematics of shapes) to uncover new patterns in clinical, energy, financial and other data sets.
They have a particular interest in healthcare and medical applications. The company was founded based on unearthing novel results for an old breast cancer data set that had already been analyzed in many ways. After a few minutes with the data set, Ayasdi’s approach identified a completely new subset of “triple negative” survivors that had elevated expression levels of genes involved in the immune system for breast cancer.
Drilling Down to Better Understand the Nuances
There are three key parts to analyzing a data set through Ayasdi’s Iris platform. First, unsupervised analysis using a mathematical (topolgical) algorithm creates a visual, fluid dynamic network, a multi-dimensional geometrical representation of the data set. Then, built-in statistical tools in Iris can be used to understand relationships between the data, easily highlighting key differences between subsets. Afterwards, a human expert can perform a “semi-supervised analysis” manipulating the structure by partitioning the data (e.g., by survival or time) to identify specific patterns of similarities and differences.
Healthcare Case Studies
Ayasdi has a number of use cases in medicine including cancer clinical trials (identifying high and low responder subgroups, finding novel biomarkers), drug toxicity, and post-traumatic stress disorder diagnosis. Through their platform, they have found specific compound side effects of drugs, correlated a particular genotype with PTSD, and classified biomarkers for cancer.
The Ayasdi Cancer Genome Atlas perfectly exemplifies their work in cancer. The Atlas characterizes tumors by somatic mutations and clinical outcomes. After creating a topological model, data analysts can identify how different population subtypes respond to particular treatments in order to design better clinical trials and eventually personalize patient treatments.
There are so many other potential life sciences applications: autoimmune disorders, neurological diseases (e.g., Alzheimer’s Disease, where research has plateaued and could use a fresh approach), and public health (e.g., obesity, diabetes). We are eager to see researchers in these areas make use of Ayasdi’s platform.
In which life science application would you like to see the Ayasdi platform speed the discovery process?