2025 marks another milestone in the evolution of cell and gene therapy (CGT). With the FDA’s recent approval of Encelto (Neurotech Pharmaceuticals) for idiopathic macular telangiectasia type 2, the total number of approved CGTs in the U.S. has risen to 44.
This figure reflects two decades of scientific ambition, translational grit, and clinical validation. But as someone who leads a biotech company built around advanced multi-omics analysis — and who works daily at the cutting edge of systems biology — I believe we’re missing something fundamental.
None of the companies behind these 44 approved therapies are publicly known to possess an in-house, fully integrated single-cell multi-omics platform. Most rely on third-party providers for sequencing and limited bioinformatics — typically confined to basic preprocessing, clustering, and a narrow form of differential gene expression analysis.
This is not a minor technical detail. It’s a systemic blind spot that threatens to limit the full potential of CGT in the years ahead.
CGT and the Need for Biological Precision
At their core, CGTs are precision interventions. Whether inserting a transgene, silencing a mutation, or transplanting modified cells, these therapies are fundamentally about engineering cellular function at a microscopic scale. Success hinges on understanding and modulating the cellular, transcriptional, epigenetic, and immunological context in which those cells operate.
Bulk data won’t cut it. Averaging across millions of cells obscures rare populations, masks dynamic cellular transitions, and ignores intercellular heterogeneity, all of which can influence therapeutic efficacy and safety.
The field has long acknowledged this in theory. Yet in practice, very few CGT developers have integrated comprehensive single-cell and multi-omics analytics into their therapeutic pipelines.
The True Complexity of CGT
Here’s what most CGT programs actually require to reduce clinical risk and increase mechanistic confidence:
- Single-cell transcriptomics to identify responsive vs. resistant cell populations
- Chromatin accessibility data (e.g. ATAC-seq) to understand regulatory network plasticity
- Multi-modal integration (RNA + protein, epigenome + transcriptome, etc.) to reveal causal biology
- Cell trajectory inference to model disease progression or therapy response
- Clonal tracing and cell fate prediction in engineered cell therapies
- Machine learning models trained on high-dimensional omics to predict off-target effects or immune responses
This level of granularity is not a luxury. It’s increasingly a clinical necessity, especially in areas like oncology, rare diseases, and regenerative medicine, where therapeutic success often hinges on resolving deep biological nuance.
The Status Quo: A Patchwork of Outsourcing and Shallow Analytics
Let’s be candid. Most CGT companies today follow this pattern:
- Outsource single-cell sequencing to a CRO or academic core
- Receive raw data and basic clustering from standard pipelines (e.g., Seurat, Scanpy)
- Generate a few UMAP plots and volcano charts for internal use or investor decks
- Conduct minimal biomarker analysis (often manually, often inconsistently)
- Lack reproducible workflows or integrated interpretation pipelines
This approach suffers from limited depth, low reproducibility, and insufficient translational value. It also leads to missed insights, especially when rare cell types, subtle signatures, or cross-modal interactions drive biological outcomes.
As someone leading Bioada, a company that builds and applies AI-powered platforms for integrative multi-omics analysis, I’ve seen firsthand how transformative it is to go beyond this paradigm.
Our Experience at Bioada
At Bioada, we’ve built a platform that goes from raw single-cell data to therapeutic and diagnostic insight — end to end. It includes:
- Automated pre-processing and rigorous QC for all omics layers
- Multi-omic cell-type discovery, not just clustering, incorporating transcriptomic, proteomic, and epigenomic features
- Dynamic modeling of cell states and transitions
- Biomarker extraction pipelines using reproducibility-validated feature selection and signal deconvolution
- Interactive visualization environments for scientists and clinicians to explore their data without coding
- AI-driven interpretation, integrating prior knowledge from public datasets, literature, and proprietary annotations
We apply this across disease contexts — including oncology, neurodegeneration, autoimmunity, and muscular dystrophies — and we’ve seen how this platform reveals mechanisms that would otherwise remain hidden.
For example, in our FSHD research, single-cell and spatial analyses allowed us to disentangle overlapping gene networks and pinpoint disease drivers previously missed in bulk studies. Our liquid biopsy biomarkers — including one-gene blood tests for breast cancer, Parkinson’s, and FSHD — emerged from integrative single-cell/multi-omics pipelines, not from traditional differential expression alone.
The Impact on CGT: From Discovery to Post-Market
Here’s where comprehensive single-cell multi-omics can shift the entire CGT development cycle:
See content credentials
Pipeline Comparison Table
This is not science fiction. These are workflows we’re actively deploying and refining, in-house, and in collaboration with partners across pharma and biotech.
The Paradigm Shift Ahead
If CGT is to fulfill its promise — durable cures, precision targeting, lower toxicity — then our analytic toolkit must evolve. As of 2025, it is no longer acceptable to approach gene and cell therapy development without high-resolution biological insight.
We must move from “sequence and guess” to “profile and predict”.
From “bulk average” to “cell-specific action”.
From disconnected data silos to fully integrated, reproducible, and interpretable pipelines.
Final Thought: CGT Without Single-Cell Insight Is Blindfolded Surgery
To those developing the next wave of CGTs: the biology you’re aiming to reprogram is deeply complex, highly contextual, and often unpredictable.
Without single-cell and multi-omics integration, you’re operating blind.
You might get lucky, but when you’re manipulating biology at this depth, luck is not a strategy.
Let’s open our eyes.