Cellarity has announced the publication of a seminal manuscript in the journal Science, which articulates a framework for the integration of advanced transcriptomic datasets and AI models to improve drug discovery.
Cellarity designs novel therapeutics for complex diseases by focusing on the interplay of pathway connections and interactions that define and modulate cellular states.
The company has built a robust discovery platform that leverages high-dimensional transcriptomics to map these interactions at single-cell resolution.
Generalisable AI models developed for the platform then link chemistry to disease biology to efficiently produce drugs that restore cellular function in diseased tissues.
The first candidate emerging from the platform, CLY-124, is under evaluation in a Phase I clinical trial for the treatment of sickle cell disease.
"We believe a comprehensive view of the cell state will help us create better therapies that can correct the foundational mechanisms of disease," said Parul Doshi, Cellarity’s Chief Data Officer.
"Our state-of-the-art platform enables us to effectively visualise this dynamic and identify novel interventions that are best suited to correct disease states."
"This publication in the prestigious journal Science describes the evaluations that have informed our platform, underscoring both the rigour and ingenuity to successfully integrate advanced transcriptomics and computational tools to enable efficient discovery of novel therapeutic candidates."
The publication in Science presents a reproducible blueprint for integrating machine learning methods into drug programs for maximum discovery potential.
The blueprint addresses numerous limitations of conventional phenotypic drug screening by employing an active, lab-in-the-loop deep learning framework powered by high-throughput transcriptomics.
By successively refining predictions based on the outcome of experiments, the framework demonstrated improved recovery of phenotypically active compounds by 13- to 17-fold over industry standard approaches.
“The drug discovery process has struggled to improve its success rates in recent decades."
"This is in part due to a conventional focus on single targets, whereas diseases are generally driven by more complex interplay than just a single gene mutation," added Jim Collins, Termeer Professor of Medical Engineering & Science, MIT, co-founder of Cellarity and co-author of the publication.
"By analysing the phenotypic connections fuelling disease pathophysiology as well as the polypharmacology considerations of early candidates, this deep learning platform offers strong potential to accelerate the pace of discovery and introduce effective new oral therapeutics for complex diseases.”
Open Source Dataset Release
Alongside its Science publication, Cellarity is releasing single-cell datasets across multiple data modalities to foster community engagement, model benchmarking and deeper insights into cell state dynamics under chemical perturbation.
The perturbational transcriptomic dataset, used to validate Cellarity’s platform, includes more than 1700 samples and 1.26 million single cells — enabling cross-cell-type drug response mapping and benchmarking of perturbation prediction models.
Additionally, Cellarity is sharing a single-cell multiomic haematopoiesis atlas integrating transcriptomics, surface receptor and chromatin accessibility data to provide a detailed view of megakaryopoiesis and erythropoiesis.
A third dataset captures a timeline of megakaryocyte (Mk) differentiation under perturbation, which can be analysed to map the trajectory of Mk maturation, interrogate time-resolved drug effects, or support model benchmarking and training.
Public analyses of these important data may yield novel insights into cellular dynamics and power new methods to accelerate industry-wide drug discovery.