When early discovery teams debate how to make better decisions, they tend to focus on new modalities, faster analytics or incremental workflow improvements. But for Shushant Jain, Director of Ncardia’s drug discovery technology group, the real leverage sits deeper: in the design of the model itself. If the biology is misaligned with the question, no amount of downstream optimization can recover meaningful insight.
Why iPSC Models Matter for Discovery Confidence
His perspective on human iPSC-derived systems comes from experience with the older staples of discovery: immortalized lines and primary cells. They were familiar, scalable, and convenient, but often poorly predictive.
“You could have great data showing your molecule is working perfectly,” he said, “but for most drug discovery programs, that translation has almost always failed. I think the failure rate is over 90 percent.”
He has seen these limitations firsthand. For example, immortalized lines lack key synaptic markers critical to the biology of many neurodegenerative diseases, and in ALS, the most fundamental disease-driving human pathways are present in patients but absent from animal models altogether. In cardiac research, they often fail to express the correct human ion channel repertoire, making them less relevant for assessing cardiac safety and efficacy.
You need to have that core biological component in your drug discovery. iPSCs restore that component, not perfectly, but far more reliably than the traditional systems he used earlier in his career.
Designing Phenotypic Assays That Reveal Biology
At Ncardia, phenotypic assays are built around biological relevance rather than experimental convenience. That means moving beyond monocultures toward multicellular systems that better reflect tissue-level interactions. It also means addressing maturity. Many diseases of interest, including cardiovascular and neurodegenerative conditions, arise in adults, while iPSC-derived cells tend to be more fetal-like. Improving maturity brings models closer to disease-relevant states.
This philosophy has had direct impact on client programs. In one case, phenotypic assays identified toxic candidates before animal testing.
“We were able to reduce the number of molecules going into an animal by about 80 percent,” Jain said.
The outcome was shorter timelines, lower costs, and a substantial reduction in animal use.
Reproducibility Through Technology and Process
One of the long-standing challenges in iPSC-based workflows is variability during differentiation. At Ncardia, this is addressed through controlled bioreactor processes, where environmental parameters such as oxygen, pH levels are continuously monitored. The result is a standardised and well-characterised cell output, with known identity and function before assays begin.
Downstream, automation plays a complementary role. Standardised handling ensures that reproducibility is maintained beyond cell production, allowing data to be compared across projects and over time.
Building Disease Models That Support Clear Interpretation
As disease models grow more sophisticated, Jain is clear that complexity alone is not the objective. What matters is whether added biological detail improves interpretability. Multicellular systems can enhance relevance, but only if they remain robust enough to produce clear, actionable data.
That principle shapes how models are developed to meet different client needs. Some programs require extensive validation of cellular identity and function, while others bring their own cells and rely on Ncardia primarily for functional assessment. When conventional validation is not feasible, models are built collaboratively, with explicit alignment on what the system can, and cannot, answer.
Across all scenarios, the underlying requirement is the same: confidence that studies are executed rigorously and that the resulting data meaningfully address the scientific question at hand.
Why Early Discovery Needs an Earlier Look at Safety
Jain sees a pattern across discovery groups. In the drive to move fast and conserve budget, safety is often pushed later into a program’s lifecycle.
“Typically, within early discovery, clients want to move fast, spend very little money, and they perhaps sometimes assess safety very late in their programs,” he said.
The consequence is predictable: candidates move forward without clarity on liabilities that could have been identified earlier. Ncardia’s approach, including its IO panel, aims to shift that thinking.
The adaptation required goes beyond assay selection. No two drug discovery programs are the same. Even with the same model or the same client, the scientific question changes, forcing experimental design and interpretation to change with it.
Where the Field Is Moving
Looking ahead, he sees two parallel paths shaping the next generation of iPSC-based discovery platforms. One involves a deeper characterization of human models and a better understanding of how translatable they truly are. The other involves helping companies unfamiliar with iPSC systems learn how to integrate them into decision-making.
Maturity remains a central focus. Many diseases of interest are adult-onset, while iPSC-derived cells still exhibit fetal-like characteristics. Jain expects this gap to continue narrowing as differentiation and maturation protocols improve.
He also points to broader industry forces driving adoption.
“It’s no longer an individual within a company asking for a more complex model,” he said. “It’s the FDA asking for them. It’s venture capitalists asking for companies to use human models to increase the chances of success in the clinic.”
For Jain, that shift represents the most significant change he has observed. Human-relevant models are no longer an experimental preference. They are becoming a baseline expectation in modern drug discovery.