How can I access your design software?
Our software is web-hosted, and can be accessed via an Application Programming Interface (API). For access through the API we need to issue you an access key, if you are interested contact us for further details. We also make our design software available for collaborative therapeutics development, which enables you to access both the awesome power of our software and our founders’ and scientists’ extensive track record in RNA Biology.

What kind of RNA Therapeutics can you design sequences for?
We design efficient sequences for any mRNA, saRNA, taRNA or circRNA based therapeutic intended to express a protein or proteins. We have most experience with mRNA Therapeutics, but have the other modalities in our portfolio. We do not design non-protein coding ASOs, or isolated siRNAs, miRNAs, or gRNAs, although we can design protein-coding sequences that are particularly responsive to small RNAs.

What kind of medical applications do you design sequences for?
We have successfully designed sequences for vaccines, protein replacement therapies for genetic diseases, neutralising antibodies, inflammation control, and growth factors. We can design sequences for any protein or application of interest.

How much more expression do you achieve compared to other approaches?
Our primary design target is therapeutic efficacy, not high expression levels. High expression levels are required for some applications such as neutralising antibodies and where we aim for high expression, we can reliably achieve expression in the top decile (based on the current state of the art, which is shifting ground - but by conducting our own research and following the literature we ensure that our approach is always up to date). In many applications, including vaccines, efficacy actually correlates poorly with expression levels and our design approach incorporates proprietary technology for controlling vaccine efficacy beyond expression levels.

What is the technical basis of your design approach?
At the heart of our design approach is a genetic algorithm that evaluates RNA sequences against a sequence-function knowledge base. Just like sequences in nature, we mature sequences from random to high-performing nucleotide composition by continually mixing and changing a pool of sequences, discarding low-performing sequences, and retaining high-performing ones.

What optimisation objectives do you consider in your design approach?
Our automated approach considers expression levels (codon usage), frameshift control (prevention of off-target peptide production), structural features, and manufacturability parameters.  In some projects we consider additional, bespoke parameters, such as RNA repeat avoidance where proteins show highly repetitive motifs, or supporting protein folding by introducing folding-supportive regions of ribosomal slow-down.

Is your technology AI?
No, our core design algorithm is not AI-based, in the sense that it does not employ neural networks or large language models. However, our approach can make selective use of AI models that relate RNA sequences to specific design objectives, such as tissue-specific regulation or interactions with LNP components.

Why does our technology excel without full AI integration?

While artificial intelligence has transformed many areas of biotechnology, we have deliberately chosen not to rely solely on AI-driven design models. Instead, our results are driven by the team's extensive research expertise, scientific insight, and years of academic excellence. There are two main advantages to this approach.
 

Firstly, all current AI-based approaches require extensive training data, meaning that model development relies heavily on the vast omics datasets available for natural mRNA sequences. The team at Jantomarna are recognised leaders in the field, and we are aware of the growing body of evidence demonstrating that therapeutic mRNAs exhibit sequence–function relationships that differ significantly from those of natural sequences. Consequently, AI models trained primarily on natural sequence data are often not well-suited to the design of high-performing therapeutic mRNAs.
 

Secondly, we aim to control and optimise all features relevant to therapeutic performance simultaneously, while retaining the flexibility to rapidly incorporate new scientific insights into our design process. 
 

We are confident that Jantomarna's approach enables higher efficacy, safety and manufacturability to be achieved in comparison to existing AI-based methods, delivering a more comprehensive, adaptable, and scientifically grounded framework for therapeutic mRNA design.