Drug discovery has long been constrained by technological limitations and an incomplete understanding of disease mechanisms and human biology. But that’s beginning to change with the rise of single-cell technologies like single-cell RNA sequencing (scRNA-seq), which enable more precise target identification, improve target validation, and support better model selection and drug mechanism analysis1. Although challenges remain, advances in scalable platforms and computational tools are making scRNA-seq a routine tool in drug discovery pipelines and enhancing the ability to translate early findings into effective therapies.
Rethinking Drug Discovery with Single-Cell Transcriptomics
One major way that single-cell transcriptomics is advancing drug discovery is by providing deeper resolution into cellular responses and identifying previously undetectable signals obscured in bulk analyses. “Many high-throughput drug screens currently rely on readouts such as bulk RNA-seq,” said Charlie Roco, Ph.D., CTO and Co-Founder of Parse Biosciences. “While bulk RNA-seq has been a step up compared to other assays that only target single or a few genes (i.e., qPCR, reporter genes, etc.), it is still missing large amounts of information per sample.” Because drug screens often involve heterogeneity from cocultures, mixed cell types, and cell cycle variation, single-cell approaches help tease apart these confounding factors and provide a clearer view of how individual cells respond to treatment.
Along with improved resolution, single-cell tools are powering large-scale, data-heavy studies in drug discovery. “We are seeing a notable shift towards using more cost-effective high-throughput technologies to do more multiplexing,” said David Peoples, Chief Financial & Business Officer at Ultima Genomics. In one example, Peoples shared that their collaboration with Tahoe Therapeutics and Parse Biosciences produced the world’s largest single-cell dataset, Tahoe-100M, in just three weeks2. The dataset captures chemical perturbations across diverse cell types. These types of studies, powered by multiplexed screening and AI-driven analysis, offer rich data for predicting off-target effects. Peoples also highlighted the Chan Zuckerberg Initiative’s Billion Cell Project and the Arc Institute’s virtual cell work as ambitious, large-scale efforts to build predictive models of cellular behavior.
Recent single-cell studies are uncovering new therapeutic targets and helping researchers predict patient-specific drug responses. Laura DeMare, Ph.D., Product Marketing at Scale Bio, explained that single-cell data can help identify drug targets expressed in specific cell types within disease-relevant tissues, which may reduce toxicity and improve clinical success, as shown in a preprint from the Teichmann Lab3. Beyond target discovery, DeMare also highlighted how single-cell data can inform patient-specific drug responses. She cited ongoing work by Drew Neavin, Ph.D., of the Garvan Institute and a winner of Scale Bio’s 100 Million Cell Challenge. Neavin developed a screening platform using 200 genetically diverse patient-derived iPSCs, which are differentiated into cardiomyocytes and exposed to various drugs to detect heart-related side effects early. This approach supports the development of safer, more personalized therapies before they reach clinical trials.
Designing a Robust Single-Cell Workflow
As scRNA-seq becomes more integrated into drug discovery, well-designed workflows are essential for producing meaningful and reproducible data. Roco emphasized that the first steps of this process begin with clear experimental goals. “Determine what the goals are and what success looks like for a particular screen or dataset and critically assess if the design is set up to achieve those goals,” he explained. This includes selecting the right cell types, defining drug or perturbation conditions, and determining how many cells per sample are needed. He noted that researchers are increasingly using computational tools to model cellular responses, describing these methods as a kind of “ChatGPT for biology.” As studies scale, particularly those involving drug or CRISPR-based perturbations, it becomes even more important to define how many perturbations to test and how much data is required to capture the biological effects with confidence.
Following the initial study design, researchers must also plan for how samples will be handled and processed across different stages of the drug discovery pipeline. DeMare emphasized that “the power of using single cell analysis as part of exploratory endpoints for clinical trials can’t be understated,” citing its ability to reveal drug mechanisms, resistance pathways, and biomarkers of response. For clinical samples, she explained that preserving biological integrity through fixation or cryopreservation is essential to ensure compatibility with downstream assays.
In preclinical studies, minimizing technical noise is key to detecting subtle transcriptional changes. Since samples from time-course experiments or drug screens are often processed on different days, DeMare advised implementing workflows that allow samples to be stored until ready for library generation. This helps reduce variability across batches and ensures more consistent, high-quality data.
Another important consideration is the sequencing platform. Peoples noted that while cost, speed, throughput, and ease of use are important across many applications, they become especially critical in large-scale single-cell studies. He explained that because most single-cell workflows involve barcode counting, sequencing quality is less of a limiting factor, shifting the emphasis toward cost, usability and integration with upstream and downstream tools. Peoples also emphasized the need for platforms that provide flexibility in batching and throughput in order to accommodate a range of experiment sizes. To address these needs, for example, Ultima recently introduced a new mode called Solaris Boost which provides a significant increase in output for short-read applications needing high scale, including single-cell. This helps researchers scale experiments, increase sequencing depth, or add other omics layers while staying within budget.
Integrating Multi-Modal Readouts
Similar to trends across the life sciences, drug discovery is shifting toward multimodal approaches, with scRNA-seq often integrated alongside other platforms to capture a broader range of biological information. DeMare pointed to a compelling example from Shift Bioscience, which used scRNA-seq to build a cellular aging atlas from over 100 donors aged 1 to 87. The team identified candidate rejuvenation genes and validated them through a large-scale genetic perturbation screen, integrating transcriptomic data with epigenetic clocks and protein biomarker predictions to assess multiomic rejuvenation effects4.
Building on this concept, Peoples described how researchers are designing large-scale multimodal studies that combine orthogonal datasets to gain deeper biological insight. In one example, he shared a collaboration with Myllia Biotechnology, where they are supporting the first platform capable of genome-wide CRISPR screening in primary human macrophages and dendritic cells. “The ability to generate ultra-high-throughput single-cell readouts at scale enables teams to link genetic perturbations with transcriptional consequences at the single-cell level, accelerating functional target validation and therapeutic hypothesis generation,” he stated.
Roco highlighted how the depth of single-cell data makes it especially valuable for AI-driven drug discovery. “Ultimately, more data that can give insight as to what biology is changing in response to perturbation will be useful for downstream AI models,” he said. With scRNA-seq capturing expression levels for approximately 20,000 genes per cell, Roco noted that the breadth of information from a single assay provides a rich foundation for training predictive models, which is one reason it's becoming a preferred input for AI in drug development.
Case Studies and Breakthroughs
Once limited to exploratory research, scRNA-seq is now at the center of some of the most innovative and actionable advances in therapeutic development. For instance, Roco highlighted two key breakthroughs from the Tahoe-100M dataset. In one case, researchers retrospectively identified off-target effects of the HIV drug Saquinavir, revealing adrenoceptor activity that had gone undetected during its time on the market. In another, the dataset uncovered novel small molecules that upregulate MHC-I expression, an immune-modulating strategy that could enhance the effectiveness of checkpoint inhibitors in cancer.
Additionally, DeMare shared that scRNA-seq helped identify predictors of therapeutic success in a clinical trial involving CAR-NKT cell products5. By comparing responders and non-responders, the analysis revealed a central memory-like T cell program associated with better expansion of the therapy in patients. This insight offers a way to assess the potential success of manufactured cell products.
The Future of scRNA-seq in Drug Discovery
With increasing maturity and broader adoption, single-cell technologies are no longer limited to exploratory work and are now becoming core components of the drug discovery pipeline. “We are at an inflection point where affordable, large-scale single-cell data generation will fuel the next wave of biologically informed drug discovery,” stated Peoples.
This development is being driven by the convergence of AI/machine learning capabilities, affordable data storage, and low-cost single-cell and sequencing technologies. “We believe this convergence is leading to a transformation in how single-cell analysis is being used in drug discovery,” added Peoples. As platforms become more scalable and multimodal datasets become more common, scRNA-seq is enabling a deeper understanding of disease biology, improving how researchers evaluate drug candidates, and accelerating the development of safer, more targeted therapies.
References
- Van de Sande B, Lee JS, Mutasa-Gottgens E, et al. Applications of single-cell RNA sequencing in drug discovery and development. Nat Rev Drug Discov. 2023;22(6):496-520. doi:10.1038/s41573-023-00688-4
- Zhang J, Ubas AA, de Borja R, et al. Tahoe-100M: A giga-scale single-cell perturbation atlas for context-dependent gene function and cellular modeling. bioRxiv. Published online May 10, 2025. doi:10.1101/2025.02.20.639398
- Dann E, Teeple E, Elmentaite R, et al. Estimating the impact of single-cell RNA sequencing of human tissues on drug target validation. medRxiv. Published online October 22, 2024. doi:10.1101/2024.04.04.24305313
- de Lima Camillo LP, Gam R, Maskalenka K, et al. A single factor for safer cellular rejuvenation. bioRxiv. Published online June 6, 2025. doi:10.1101/2025.06.05.657370
- Heczey A, Xu X, Courtney AN, et al. Anti-GD2 CAR-NKT cells in relapsed or refractory neuroblastoma: updated phase 1 trial interim results. Nat Med. 2023;29(6):1379-1388. doi:10.1038/s41591-023-02363-y