Tejal Somvanshi
Harvard Medical School researchers have created a new AI tool that could transform how doctors treat Parkinson’s, Alzheimer’s, and even cancer. The system, called PDGrapher, works by finding the exact genes and drug combinations that might push diseased cells back to a healthy state.
Unlike traditional drug discovery methods that test one target at a time, PDGrapher looks at the bigger picture. It maps the complex connections between genes and proteins to identify multiple targets that together could reverse disease.
“Traditional drug discovery resembles tasting hundreds of prepared dishes to find one that happens to taste perfect,” explained Marinka Zitnik, associate professor of biomedical informatics at Harvard. “PDGrapher works like a master chef who understands what they want the dish to be and exactly how to combine ingredients to achieve the desired flavor.”
This approach could help solve a major problem in medicine: some diseases are too complex to treat by targeting just one protein. Cancer cells, for example, often develop resistance to single-target drugs.
The team tested PDGrapher on 19 different datasets covering 11 types of cancer. The results were impressive – the AI ranked the correct therapeutic targets up to 35% higher than other models and delivered results up to 25 times faster.
One key success was identifying KDR (also known as VEGFR2) as a promising target for non-small cell lung cancer. The system also highlighted TOP2A, an enzyme targeted by some existing chemotherapies, as potentially useful for preventing cancer spread.
The researchers are already putting PDGrapher to work on brain diseases. They’re studying how cells behave in Parkinson’s and Alzheimer’s, looking for genes that could help restore healthy function. They’ve also partnered with Massachusetts General Hospital to study X-linked Dystonia-Parkinsonism, a rare inherited disorder.
What makes PDGrapher different from previous AI tools is its focus on disease processes rather than single drug targets. It aims to understand the underlying cellular dysfunction and find ways to correct it.
“Our ultimate goal is to create a clear road map of possible ways to reverse disease at the cellular level,” Zitnik said.
Similar Posts
The project received support from both government agencies (NIH, NSF, ARPA-H) and private organizations (Chan Zuckerberg Initiative, Gates Foundation), along with partnerships with pharmaceutical companies including AstraZeneca, Roche, Sanofi, and Pfizer.
Looking ahead, the team envisions a future where doctors could analyze an individual patient’s cellular profile and design personalized treatment combinations. This could be especially valuable for complex conditions where standard treatments often fail.
PDGrapher is available through GitHub as a free research tool, allowing scientists worldwide to use it in their own work. This open approach could speed up research in areas that have long challenged traditional drug discovery methods.
The Harvard team’s work joins other recent AI breakthroughs in medicine, such as Google DeepMind’s AlphaFold, which predicts protein structures. Together, these tools suggest a future where AI increasingly helps solve biology’s most complex puzzles.
Harvard notes that the research was partially funded by federal grants, and the future of such federally funded work now faces uncertainty due to recent government decisions affecting grant funding across Harvard University.
