Connecting molecular profiling to real-world patient experiences
The majority of new therapies submitted for Food and Drug Administration (FDA) approval in the United States fail due to a lack of efficacy. I believe this stems from several structrual problems in biomedical research. First, hypothesis generation for drug development focuses too narrowly on model systems and hyper controlled scenarios, not on real-world patient tissues. Second, biological and biomedical data are extremely siloed and static-most datasets are only used for one or two studies, despite their expense, and they are difficult to access and work with.
My work at Enable Medicine, a biotechnology company, focused on breaking down this status quo through the following big-picture aims:
- Building access to real-world human tissue samples linked to deidentified electronic health records
- Profiling disease in these samples at a large scale, using spatial methods that preserve the native tissue context of cells and molecules
- Designing computational and AI/ML tools that facilitate data integration