Artificial intelligence uncovers previously unknown cellular signatures of Parkinson’s disease
A study was recently published in Nature Communications that presents a novel automated cell culture platform, which researchers utilized to uncover previously unknown cellular signatures of Parkinson’s disease. The researchers suggest this could be employed to discover more effective drugs for a range of diseases.
The platform utilizes artificial intelligence (AI) technology to study images of patient cells. Scientists from the New York Stem Cell Foundation (NYSCF; NY, USA) reported that they have identified the specific cellular hallmarks of Parkinson’s disease, which could be utilized to discover better treatment for the condition.
The discovery was made by creating and profiling over one million images of skin cells from a group of 91 patients and healthy controls. The research team utilized the platform to isolate and expand fibroblasts from skin punch biopsy samples and labelled different parts of the cells to create thousands of high-content optical microscopy images.
Researchers then put the images into an unbiased, AI-driven image analysis system, which identified particular features that were specific to patient cells. The research team suggest that this could be utilized to determine them from healthy controls.
The research team proposed that the discovery of novel disease signatures may be valuable for diagnostics and drug discovery, and could even reveal previously indistinguishable distinctions between patients.
The Parkinson’s disease signatures that have been identified could also be utilized as a basis for conducting drug screens on patient cells. The study yielded the largest known cell painting dataset, which is available for use by the research community.
Susan Solomon, CEO of NYSCF, stated: “traditional drug discovery isn’t working very well, particularly for complex diseases like Parkinson’s…the robotic technology NYSCF has built allows us to generate vast amounts of data from large populations of patients, and discover new signatures of disease as an entirely new basis for discovering drugs that actually work.”
Due to the fact that the platform requires easily accessible skin cells from patients, researchers state that it could be applied to other cell types, including derivatives of iPSCs that NYSCF produces to model a variety of diseases. The team are confident that the platform will open therapeutic avenues for many other diseases.
Source: Solomon S, Schiff L, Chen Y et al. Integrating deep learning and unbiased automated high-content screening to identify complex disease signatures in human fibroblasts. Nat. Commun. 13(1590) (2022).