Method to identifying drug targets from single-cell perturbation screens

This technology is an integrated experimental and computational method that identifies drug targets by combining CRISPR perturbation screens, single-cell transcriptomic readouts, and a deconvolution model to support drug development.

Unmet Need: Reliable identification of drug targets and mechanisms of action

Many drugs, including some advanced into clinical trials, act through targets that differ from their reported mechanism, complicating safety assessment and limiting therapeutic efficiency. Genetic perturbation screens combined with drug treatments are widely used to validate drug targets, but standard analyses rely on coarse readouts like cell viability or sgRNA enrichment, discarding the high-resolution information now available from single-cell transcriptomics. Drug developers need analytical methods that can directly extract drug target signals from the rich transcriptomic data generated in modern perturbation and drug screens, while quantifying confidence in result.

The Technology: A computational and experimental method to identify drug targets from single-cell perturbation and drug screens

This technology is computational framework for determining the targets of a drug from single-cell transcriptomic readouts of CRISPR perturbation and drug screens. Cells receive perturbation constructs carrying a guide RNA, a perturbation barcode, and a reverse-transcription handle. The perturbed and drug-treated cells are profiled by single-cell RNA sequencing. The computational model then examines the resulting gene expression data using a latent factor model to deconvolute shared baseline variation or drug-specific changes for each cell types. A Bayesian regression approach then explains drug treatment effects as combinations of perturbation effects to identify candidate drug targets. The framework outputs a ranked set of candidate drug targets along with their associated probabilities, providing interpretable target predictions from existing single-cell screen data.

This technology has been validated on published single-cell perturbation and drug screen datasets, achieving a median sensitivity of 92% and median precision of 99% in classifying cellular perturbations and correctly identifying BCR as the highest-probability target of Bcr-Abl inhibitors across datasets generated with different sequencing methods.

Applications:

• Drug development
• Deconvolution, validation, and discovery of drug targets
• Drug off-target screening
• Drug safety assessment
• Identification of biomarkers for clinical trials
• Drug mechanism of action research
• Research tool for analysis of perturbed genes and pathways in disease

Advantages:

• Computationally scalable to large datasets
• Integrates information from both drug screening and perturbation studies
• Resolves targets from high-resolution single-cell data, not coarse viability or marker readouts
• Calibrates confidence score for every candidate target
• Transferable across different sequencing methods and datasets

Lead Inventor:

David Knowles, Ph.D.

Patent Information:

Patent Pending

Related Publications:

Tech Ventures Reference:

• Licensing Contact: Joan Martinez