close
Neuroscience

Artificial Intelligence and Robotics Discover Parkinson’s Disease’s Hidden Signatures

A study published today in Nature Communications describes a new platform for identifying disease-related cellular signatures that combines robotic tools for analyzing patient cells with artificial intelligence algorithms for picture analysis.

Scientists at the NYSCF Research Institute collaborated with Google Research to successfully identify new cellular hallmarks of Parkinson’s disease by creating and profiling over a million images of skin cells from a cohort of 91 patients and healthy controls using their automated cell culture platform.

“Traditional drug development isn’t performing very well,” said NYSCF CEO Susan L. Solomon, JD. “This is especially true for complex disorders like Parkinson’s.” “The robotic technology that NYSCF has developed enables us to produce massive volumes of data from large populations of patients and uncover novel disease signatures as an altogether new basis for discovering medications that actually work.”

“This is an ideal demonstration of the power of artificial intelligence for disease research, We have had a very productive collaboration with NYSCF, especially because their advanced robotic systems create reproducible data that can yield reliable insights.”

Marc Berndl, Software Engineer at Google Research.

“This is excellent proof of the capacity of artificial intelligence for illness research,” said Marc Berndl, a Google Research Software Engineer. “Our work with NYSCF has been quite fruitful, especially because their advanced robotic systems generate reproducible data that can yield accurate insights.”

Bringing Artificial Intelligence and Automation Together

The study took advantage of the NYSCF’s huge collection of patient cells and cutting-edge robotic equipment—the NYSCF Global Stem Cell Array®—to profile images of millions of cells from 91 Parkinson’s patients and healthy controls. The Array® was used by scientists to extract and expand skin cells called fibroblasts from skin punch biopsy samples, label distinct regions of these cells using a technique called Cell Painting, and generate hundreds of high-content optical microscopy images.

The images were then fed into an unbiased, artificial intelligence–powered image processing pipeline, which identified visual characteristics unique to sick cells that could be used to distinguish them from healthy controls.

“These artificial intelligence algorithms can discover what patient cells have in common that would otherwise be invisible,” stated Samuel J. Yang, Research Scientist at Google Research. It’s also crucial to note that the algorithms are unbiased—they don’t rely on existing knowledge or prejudices about Parkinson’s disease, allowing us to find whole new disease markers.

The significant failure rates of recent clinical trials for medications based on specific disease targets and pathways thought to be disease drivers highlight the need for new Parkinson’s signatures. The finding of these novel illness signatures utilizing unbiased approaches, particularly across patient groups, offers importance for diagnostics and medication research and may potentially show new distinctions between patients.

“We were able to differentiate between images of patient cells and healthy controls, as well as between distinct subtypes of the disease,” said Bjarki Johannesson, PhD, a NYSCF Senior Investigator on the project. “We could even predict which donor a sample of cells came from fairly well.”

Drug Discovery Applications

The team’s discovery of Parkinson’s disease markers can now be utilized to run pharmacological screenings on patient cells to see whether medications can reverse these traits. The project also produces the largest known cell painting dataset (48TB) as a community resource, which is freely available to the research community.

Notably, the technology is disease-agnostic, requiring only patients’ easily available skin cells. It can also be applied to other cell types, such as induced pluripotent stem cell derivatives created by NYSCF to represent a variety of illnesses. As a result, the researchers are optimistic that this platform will open up new treatment paths for many diseases where traditional drug development has failed.

“This is the first tool that has effectively identified disease traits with this level of precision and sensitivity,” stated Daniel Paull, PhD, Senior Vice President of Discovery and Platform Development at NYSCF. “Its ability to identify patient subgroups has significant implications for precision medicine and drug development across a wide range of difficult-to-treat disorders.”

Topic : Article