Nazli Azimi, Co-Founder and CEONazli Azimi, co-founder and CEO, had already spent years in biotech, first as a scientist and then as an entrepreneur. Her previous two companies had taken immunology programs into Phase II clinical trials before being acquired. She was not looking for another company to build. Then she met Hani Goodarzi, scientific co-founder, who was looking to turn a lifetime of work in RNA biology and machine learning (ML) into something that could move beyond the laboratory and into drug development.
Goodarzi had spent his life studying RNA biology and was among the early researchers to incorporate ML into the field, beginning in graduate school and continuing through his postdoctoral research and academic appointments at UCSF and the Arc Institute. For Azimi, what stood out was that he understood both sides of a problem most companies approached from only one direction. He understood the complexity of RNA biology and the AI and ML needed to model it and spoke both languages fluently.
That distinction mattered. For Azimi, AI could only be as powerful as its understanding of RNA biology, whose structure, regulatory elements, and behavior across cells and tissues shape how a medicine works.
Azimi brought something equally important to the partnership, years of experience turning biological discoveries into medicines. Together, they saw a way to bring those two disciplines together.
That became Therna Biosciences. For Azimi, the combination was not simply interesting. It was necessary.
Today, the company has built RNA-Logix™, a biology-first RNA intelligence platform, grounded in a proprietary RNA knowledge base and designed to produce programmable RNA medicines by combining deep RNA biology, proprietary experimental data, and ML to dramatically shorten the path from concept to therapeutic candidate.
Turning RNA into a Programmable Medicine
RNA-Logix™ was built around a simple idea. Every disease presents a different biological challenge, so every RNA medicine should be designed with properties tailored to that challenge instead of relying on a one-size-fits-all approach.
The platform combines deep RNA biology, large-scale experimental data, and ML to create programmable RNA therapeutics. Researchers can design RNA molecules with characteristics tailored to a specific disease, including tissue targeting, durability, and controlled protein expression. Rather than optimizing one characteristic at a time, the platform evaluates how they interact, allowing multiple biological objectives to be addressed simultaneously.
-3.jpg)
Hani Goodarzi, Scientific Co-Founder Unlike conventional AI models that are designed primarily to predict outcomes, RNA-Logix™ is built to reason across these interconnected biological variables. By drawing on its integrated RNA knowledge base, the platform can come up with hypothesis and explain why a particular RNA design is recommended for a given disease, test it against its own vast data set recursively and deliver a refined proposal for an RNA design.
This ability to reason across complex biological systems has enabled RNA-Logix™ not only to design therapeutic candidates but also to reveal previously unknown aspects of RNA biology that were later confirmed experimentally.
The same biological foundation extends across multiple RNA modalities. The platform supports the design of messenger RNA therapeutics, where RNA itself serves as the medicine, as well as antisense oligonucleotides and small interfering RNA therapies as well small molecules that target RNA regulatory elements to increase, decrease, or fine-tune protein production. Rather than building separate discovery engines for different modalities, Therna Biosciences applies a single biological knowledge base across them, giving partners the flexibility to pursue diverse therapeutic strategies using one integrated platform.
Equally important is how the platform continues to evolve. Every computational prediction is validated experimentally before the results are fed back into the system, creating a continuous learning cycle between the laboratory and the computational models.
We’ve developed proprietary experimental methods that generate up to 50 times more high-quality biological data than conventional approaches.
According to Azimi, the quality of an AI model begins long before the first algorithm is trained. It starts with asking the right biological questions and designing experiments that faithfully capture the complexity of RNA and the biological systems surrounding it. That philosophy has shaped Therna Biosciences’ entire discovery process, from proprietary experimental methods to the datasets used to train RNA-Logix™.
“We’ve developed proprietary experimental methods that generate order of magnitudes more high-quality data in complex biological systems than conventional approaches,” says Azimi. “That growing body of proprietary data strengthens the platform over time, enabling increasingly reliable recommendations while continuously expanding its understanding of RNA biology.”
That combination of biological reasoning, proprietary experimental data, and continuous model refinement has fundamentally changed the pace of early-stage RNA design. According to Azimi, what has traditionally taken years and decades of iterative experimentation to discover and design medicines can now be compressed into a process measured in weeks, with RNA-Logix™ capable of generating new design recommendations within days before they move into experimental validation.
Unlocking New Therapeutic Possibilities
One of RNA-Logix™'s biggest strengths lies in its ability to identify opportunities that conventional discovery approaches might overlook. Rather than scanning RNA molecules through traditional trial-and-error screening, the platform analyses the biological context surrounding an RNA transcript to determine where therapeutic intervention is most likely to succeed. This allows researchers to identify regions that are more likely to deliver meaningful biological outcomes before extensive laboratory screening begins.
The impact of this approach is particularly evident in the development of antisense oligonucleotides (ASOs) and small interfering RNA (siRNA) therapies. Traditionally, researchers design numerous oligonucleotides across an RNA transcript and test each one experimentally to determine which regions are effective. Therna Biosciences takes a different route. By combining its RNA knowledge base with biological reasoning, RNA-Logix™ recommends specific regions within an RNA molecule that are most suitable for therapeutic targeting, significantly narrowing the search space and allowing researchers to focus their efforts on the most promising candidates.
More importantly, the platform is not limited to identifying known targets. According to Azimi, RNA-Logix™ has repeatedly uncovered previously uncharacterized regulatory regions within RNA molecules that had not been associated with therapeutic intervention. The company then synthesized ASOs against those regions and validated the predictions through laboratory studies and animal models, demonstrating that these newly identified sites could indeed be targeted effectively. For partners, this represents more than a faster discovery process. It creates opportunities to pursue novel therapeutic targets that may have remained hidden using conventional screening methods.
-
Built on years of experimental research, rna-logix™ brings together information on rna structure, regulatory elements, tissue specificity, durability, protein expression, manufacturability, safety, and cellular context into a unified framework, to represent the complexity of the biological systems.
More importantly, the platform is not limited to identifying known targets. According to Azimi, RNA-Logix™ has repeatedly uncovered previously uncharacterized regulatory regions within RNA molecules that had not been associated with therapeutic intervention. The company then synthesized ASOs against those regions and validated the predictions through laboratory studies and animal models, demonstrating that these newly identified sites could indeed be targeted effectively. For partners, this represents more than a faster discovery process. It creates opportunities to pursue novel therapeutic targets that may have remained hidden using conventional screening methods.
From Platform to Proof
Therna Biosciences is pursuing areas of unmet medical need where RNA-Logix platform can provide differentiated medicines for large population. However, one area where it believes RNA-Logix™ could have an immediate and big impact is personalized medicine. Patients with ultra-rare genetic disorders often have limited or no treatment options because developing therapies for a single patient or a very small population rarely fits the economics of traditional drug development. Therna Biosciences believes programmable RNA can help change that equation.
-6.jpg)
"We've already demonstrated that we can design individualized RNA medicines for a single patient with severe lung fibrosis and rare neurological disorders within months. Those programs have since moved into preclinical development," says Azimi.
That speed is made possible in part by RNA-Logix™'s ability to generate high-quality RNA designs in a zero-shot setting, allowing researchers to begin with drug-like candidates instead of refining molecules through repeated cycles of computational optimization.
Working alongside organizations such as Charles River, those programs progressed into preclinical development, demonstrating how a faster design process could help accelerate potential treatment options for patients with urgent medical needs. While these programs are still in development, they highlight the practical applications of RNA-Logix™ beyond conventional discovery timelines.
Looking Beyond Today's RNA Medicines
Despite the rapid progress of RNA therapeutics, delivery continues to be one of the biggest hurdles facing the industry. Designing an effective RNA medicine is only part of the equation. Ensuring that it reaches the right tissue safely and efficiently remains equally critical.
-7.jpg)
Therna Biosciences is not attempting to solve the industry's delivery challenge on its own. Instead, the company's focus remains on engineering the RNA cargo itself while leveraging advances in delivery technologies being developed across the broader biotechnology ecosystem. This includes a collaboration with leading investigators at UPenn in RNA delivery technology. As those delivery systems continue to evolve, Therna Biosciences expects to integrate its RNA design capabilities with them, creating therapies that combine optimized cargo with increasingly effective delivery mechanisms.
Looking further ahead, the company's ambitions extend beyond designing individual RNA medicines. Azimi envisions a future where increasingly sophisticated biological models can capture the interconnected nature of human biology, enabling researchers to better understand how therapies influence not only their intended targets but also the wider biological system. It is a long-term vision that reflects the company's broader philosophy: meaningful advances in drug discovery begin with a deeper understanding of biology.
As the RNA therapeutics landscape continues to evolve, Therna Biosciences is positioning itself as a leader in the field, helping researchers design smarter RNA medicines from the very beginning. By bringing together deep RNA biology, proprietary experimental research, and computational design within a single discovery framework, the company is creating new possibilities for how RNA therapeutics are discovered, engineered, and translated into the next generation of precision medicines.
Selecting Programmable RNA Therapeutics through Biological Fidelity
A model can generate an RNA sequence quickly and still leave a development team with months of avoidable laboratory work. The cost appears later, when a promising design fails to preserve the intended behavior in the relevant cell type or animal system. RNA function is shaped by sequence context, regulatory architecture, molecular structure and expression kinetics. Buying decisions therefore hinge less on raw generation speed than on whether the platform captures those dependencies before candidate selection.
Single-property optimization creates a familiar trap. Increasing persistence may alter protein output, while improving expression can weaken tissue restriction. A credible system must reason across linked biological properties rather than optimize one variable in isolation. Executives should examine how the model represents cell context and whether its predictions account for the surrounding transcript environment. A platform built around narrow sequence scoring may produce attractive rankings without explaining why a molecule should perform in a specific tissue.
Training data deserves equal scrutiny. Public datasets are useful for pre-training, but they rarely provide the consistency needed for therapeutic design. Experimental methods, assay conditions, sample handling and biological systems can differ enough to distort model performance. The stronger approach generates proprietary data under controlled conditions, tests predictions in relevant in vitro and in vivo settings, documents failures and returns those findings to the model. This closed loop matters because every failed design should sharpen the next design round rather than remain an isolated laboratory result.
Reported model accuracy can also hide uneven performance across tissues or sequence classes. Review should cover data lineage, assay reproducibility, model versioning and the threshold for moving a design into animal testing. Buyers also need to know whether partner-generated evidence improves future models or remains separated from the core learning system.
Modality coverage requires more than a shared software interface. Designing mRNA therapeutics involves programming durability, tissue restriction, expression level and protein output, while oligonucleotide programs depend on identifying accessible and biologically meaningful regions of a transcript. ASO or siRNA design cannot be reduced to faster tiling. The platform should narrow the search space using cellular context, providing a rationale that experimental teams can test. Buyers should also assess whether the same knowledge base can support distinct therapeutic approaches without forcing them into one generalized model.
Translation speed becomes meaningful only when the platform can move from prediction to a testable molecule through a defined handoff. Internal validation capacity reduces interpretation gaps between computational teams and bench scientists. It also gives management a clearer view of candidate rationale, supporting evidence, remaining uncertainty and the work required before preclinical development. Clear ownership of generated data matters as well, particularly in partnerships where experimental findings may influence later design cycles.
Therna Biosciences warrants consideration as a premier choice through RNA-Logix, which links RNA foundation models to a proprietary lab-in-the-loop system. Experimental data from in-house validation feeds subsequent design cycles, keeping computational output tied to measurable biology. Its logic treats RNA behavior as an interconnected system and supports programmable mRNA medicines and oligonucleotide programs. For ASO and siRNA designs, context-aware prediction can narrow target regions before laboratory testing. Buyers prioritizing biological fidelity, traceable validation, modality-specific design and a shorter route from sequence proposal to tested molecule should view Therna as a focused option.
...Read more
Programmable RNA Therapeutics Company Info
Q1
What Does Programmable RNA Therapeutics Development Involve?
Programmable RNA therapeutics development involves designing RNA-based medicines with properties suited to a particular biological challenge rather than applying the same molecular approach to every disease. Researchers may consider factors such as tissue targeting, durability, protein expression, safety and manufacturability together because changes in one property can affect another. The goal is to create RNA candidates whose biological behavior is deliberately engineered for the intended therapeutic application.
Q2
How Does Therna Biosciences Approach Programmable RNA Therapeutics Development?
Therna Biosciences approaches programmable RNA therapeutics development through RNA-Logix™, a biology-first RNA intelligence platform built on a proprietary RNA knowledge base. The platform combines RNA biology, proprietary experimental data and machine learning to evaluate interconnected biological variables during RNA design. It can also support multiple therapeutic modalities, including messenger RNA, antisense oligonucleotides and small interfering RNA, using a shared biological knowledge foundation.
Q3
Why Is Biological Context Important When Designing RNA Medicines?
Biological context is central to programmable RNA therapeutics development because an RNA sequence does not operate independently of the cell, tissue or disease environment surrounding it. Structure, regulatory elements and cellular behavior can all influence therapeutic performance. Considering these relationships earlier can help researchers identify promising intervention points and avoid spending extensive experimental effort on regions that are less likely to produce a meaningful biological effect.
Q4
How Can Programmable RNA Therapeutics Development Improve the Discovery Process?
Programmable RNA therapeutics development can narrow the path from a broad biological question to a therapeutic candidate by examining multiple design objectives together. Instead of testing large numbers of molecules through repeated trial-and-error cycles, researchers can use computational and experimental knowledge to prioritize candidates or target regions for further validation. The approach still depends on laboratory testing, but better-informed starting points can reduce unnecessary iterations and focus experimental work on more relevant possibilities.
Q5
What Makes Therna Biosciences' Development Model Different?
A defining part of Therna Biosciences' programmable RNA therapeutics development model is the continuous connection between computational prediction and experimental validation. Predictions generated through RNA-Logix™ are tested experimentally, and the resulting data is used to strengthen subsequent design recommendations. The company also uses proprietary experimental methods to generate large volumes of biological data, while its platform has identified previously uncharacterized RNA regulatory regions that were subsequently investigated through laboratory and animal studies.
Q6
Where Could Programmable RNA Medicines Have the Greatest Impact?
Programmable RNA therapeutics development may be particularly valuable where conventional drug discovery timelines or economics make treatment development difficult. Therna Biosciences has explored individualized RNA medicines for ultrarare conditions, including programs involving severe lung fibrosis and rare neurological disorders that moved into preclinical development. More broadly, the approach could expand the ability to tailor RNA medicines to specific biological requirements while allowing researchers to pursue therapeutic strategies across different RNA modalities.
