NovaEngine
- AI design system
- 4th-gen generative AI
NovaEngine™ is a generalist RNA design system built to address the most complex molecular targets across disease areas. The partner model puts it directly in your hands.
Novarna’s capabilities have been built iteratively, validated at each stage before advancing. Every generation of NovaEngine adds a new design capability, and the data it generates feeds the next.
Outperforms Evo2 across all 6 RNA classes (p < 0.01). NovaFM enables faster, more efficient predictions than models trained on trillions of nucleotides.
Generative design across circRNA sponge delivery, miRNA degraders, TDcircRNA protein degradation, and aptamers — all confirmed in wet lab conditions.
Integrates aptamer design and experimental data into the AI model — turning aptamer screening into AI-guided engineering.
NovaEngine and NovaLab™ form a closed feedback loop that gets smarter with every experiment. The lab generates proprietary data that retrains the AI — and the AI designs better candidates for the lab to validate.
Traditional platforms screen sequences from existing libraries. NovaEngine generates novel RNA from scratch — fewer candidates, faster cycles, lower cost per program.
NovaLab generates proprietary experimental data that retrains NovaEngine with every cycle. No external dataset can replicate it. Every program makes the platform more accurate.
Partners come to Novarna with a biological target and a disease indication. We identify the right non-coding RNA modality from our platform: microRNA sponges, aptamers, aptamer drug conjugates, targeted protein degradation. The platform adapts to the therapeutic challenge, not the other way around.
Most RNA therapeutics are designed to do a single thing. Novarna designs molecules that do several — packaged inside a single circular RNA structure. This is possible because RNA is structurally modular: individual functional units can be linked together, each folding and operating independently. Nature already works this way. Until now, designing multi-module RNA structures reliably has been beyond reach. NovaEngine makes it an engineering problem rather than a guessing game.
NovaEngine designed circular RNA with multiple therapeutic functions in one structure. In lab testing, those designs found and eliminated a disease-driving protein inside cancer cells. Precision design at the molecular level.
RNA’s inherent modularity is what makes this possible. Unlike proteins or synthetic small molecules, RNA structures can be engineered as independent building blocks, linked together into one deliverable molecule. Each function operates without interfering with the next. And because RNA is a natural molecule that the cell already knows how to use, it interfaces with biology in a way a foreign synthetic compound never can. NovaEngine is the design tool that makes this engineering possible.
Novarna designs RNA that intercepts the signals cancer cells use to grow and resist treatment, validated against a commercial benchmark in ovarian cancer models at the University of Michigan. The designs outperformed a commercially available treatment.
Best fit: programs targeting specific cancer-driving genes or proteins.
Confirmed PoCRare disease programs require highly specific solutions — and rarely have the scale for years of trial-and-error. Novarna designs directly to the target. Discovery time drops from years to months. Every design is built for that target specifically, not adapted from a library of approximate sequences.
Best fit: programs where existing tools have stalled or the target is too specific to screen by hand.
In developmentCardiovascular programs need RNA candidates that are stable in the body and tissue-specific. Novarna designs to those specifications from day one — reducing the time spent optimizing later.
Best fit: programs targeting lipid pathways, cardiac inflammation, or fibrosis.
In developmentThe brain is difficult to reach and unforgiving of imprecision. Novarna designs RNA candidates built specifically for that environment — structurally optimized for a single target, not a broad category.
Best fit: programs targeting neurodegeneration, gene silencing in the CNS, or intracellular proteins.
In developmentTraditional aptamer screening takes months of iterative lab work. Novarna generates and ranks candidates computationally first, then confirms the best in our wet lab — delivering leads in weeks, not months. The modality is adaptable, covering aptamers, aptamer drug conjugates, or combinations designed for your specific target.
Best fit: programs needing a high-affinity binder to a specific protein target.
In developmentNovaEngine is a generalist design system. If you have a target, we’d like to evaluate fit.
Novarna embeds our design system directly into your research workflow — from target definition through candidate delivery — through a structured partnership model that scales with your program’s needs.
↺ Lab data retrains NovaEngine — every experiment improves the model
Partner works with NovaLab to define target parameters, therapeutic goals, and design constraints.
NovaEngine generates novel RNA sequences optimized to your target — not screened from a library, but designed from scratch.
NovaLab synthesizes candidates and runs preliminary screening, or works with your lab to test in your existing assays.
Experimental results retrain the AI — every iteration improves model accuracy for your program and future programs.
You license the AI-designed candidates for further development and commercialization, with milestone and royalty structures aligned to your program.
Our science team works directly with research partners to evaluate target fit and scope a program.