09-15-2025 
|
Getting your Trinity Audio player ready...
|
Researchers at Harvard Medical School in the United States revealed that they have designed an artificial intelligence (AI) model capable of identifying treatments that can restore diseased cells, in a move that could “reshape drug discovery.”
The medical school of Harvard University recently announced that it is has been developing a new AI tool that can accurately identify multiple drivers of disease in cells and predict effective therapies, potentially helping in the treatment of cancer and degenerative brain diseases such as Alzheimer’s and Parkinson’s.
The work, which was supported in part by federal funding, attempts to move away from traditional drug discovery approaches that look for and target single sources of cell dysfunction and instead aim to address the underlying disease processes.
“Unlike traditional approaches that typically test one protein target or drug at a time in hopes of identifying an effective treatment, the new model, called PDGrapher and available for free, focuses on multiple drivers of disease and identifies the genes most likely to revert diseased cells back to healthy function,” stated a Harvard Medical School report on the research.
It added that “the new AI model sets the stage for better drug discovery and could lead to better individualized therapies.”
The research team currently uses the model to tackle brain diseases such as Parkinson’s and Alzheimer’s by seeing how cells behave in disease and identifying genes that could help restore them to health.
AI changing the game
PDGrapher is a type of AI tool called a “graph neural network” designed to map the relationship between various genes, proteins, and signaling pathways inside cells and predict the best combination of therapies that would help repair damaged or diseased cells.
According to the Harvard report, by “zeroing in on the targets most likely to reverse disease,” the new approach could speed up drug discovery and design, as well as unlock therapies for conditions that have proven elusive to medicine.
“Traditional drug discovery resembles tasting hundreds of prepared dishes to find one that happens to taste perfect,” said the study’s senior author, Marinka Zitnik, associate professor of biomedical informatics in the Blavatnik Institute at Harvard Medical School. “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.”
Specifically, while traditional drug discovery approaches tend to focus on activating or inhibiting a single protein, PDGrapher looks at the “bigger picture” to find compounds that can reverse signs of disease in cells.
As explained in the publication, PDGrapher first points to parts of the cell that might be driving disease. Next, it simulates what happens if these parts are turned off or dialed down. Finally, it offers an answer as to whether a diseased cell would occur if certain targets are hit.
“Instead of testing every possible recipe, PDGrapher asks: ‘Which mix of ingredients will turn this bland or overly salty dish into a perfectly balanced meal?’” said Zitnik.The study
The report outlined how the researchers trained PDGrapher on a dataset of diseased cells before and after treatment, “so that it could figure out which genes to target to shift cells from a diseased state to a healthy one.”
They then tested it on 19 datasets across 11 types of cancer, asking the AI tool to predict various treatment options for cell samples it had not seen before and for cancer types it had not encountered.
The results were extremely promising, PDGrapher “accurately predicted drug targets already known to work but that were deliberately excluded during training to ensure the model did not simply recall the right answers,” as well as identifying additional candidates supported by emerging evidence.
When compared to similar tools and comparable AI approaches, the model showed “superior accuracy and efficiency,” ranking the correct therapeutic targets up to 35 percent higher than other models did and delivering results up to 25 times faster.
Implications
When dealing with the treatment of serious diseases, the stakes can be life and death. Any new approach to improving treatment, such as optimizing the way new drugs are designed, could have huge impact on peoples’ lives, as much as on medicine more broadly.
In this respect, the research indicates that PDGrapher could speed up the testing of ideas and allow researchers to focus on fewer promising targets.
This could be particularly important in treating complex diseases, such as cancer, “in which tumors can outsmart drugs that hit just one target” — by identifying multiple targets involved in a disease, PDGrapher could help circumvent this problem.
The Harvard report also noted that the researchers were hopeful that, after careful testing to validate the model, it could one day help design individualized treatment plans for different patients.
“Our ultimate goal is to create a clear road map of possible ways to reverse disease at the cellular level,” said Zitnik.
In order for artificial intelligence (AI) to work right within the law and thrive in the face of growing challenges, it needs to integrate an enterprise blockchain system that ensures data input quality and ownership—allowing it to keep data safe while also guaranteeing the immutability of data. Check out CoinGeek’s coverage on this emerging tech to learn more why Enterprise blockchain will be the backbone of AI.
Watch: Adaptable blockchain system to tackle real-world problems
Recommended for you
