Poster Presentation Australasian RNA Biology and Biotechnology Association 2026 Conference

An Integrated Design–Build–Test Workflow for Computational Engineering and Functional Validation of Synthetic mRNAs  (#146)

Zaka Yuen 1 2 3 4 , Agin Ravindran 1 2 3 5 , Susannah Brown 1 3 , Nicole Thomsen 1 3 , Amee George 1 3 6 , Gaetan Burgio 1 3 , Eduardo Eyras 1 2 3 5
  1. The Shine-Dalgarno Centre for RNA Innovation, , The John Curtin School of Medical Research, Canberra city, Australian Capital Territory, Australia
  2. EMBL Australia Partner Laboratory Network, The Australian National University, Canberra City, ACT, Australia
  3. The National Thereapeutic mRNA platform, The Australian National University, Canberra, ACT, Australia
  4. The Centre for Computational Biomedical Sciences, The John Curtin School of Medical Research, Australian National University, Canberra, Canberra, ACT, Australia
  5. The Centre for Computational Biomedical Sciences, The John Curtin School of Medical Research, Canberra city, Australian Capital Territory, Australia
  6. The ANU Centre for Therapeutic Discovery, The John Curtin School of Medical Research, Australian National University, Canberra, Canberra, ACT, Australia

Rational design of therapeutic mRNAs requires integration of computational sequence engineering, molecular manufacturing, and functional validation, but these components are often performed independently, limiting iterative optimisation and slowing translation into therapeutic applications. Here, we present an integrated Design-Build-Test framework that combines computational mRNA design, standardised molecular manufacturing, and quantitative functional validation into a streamlined, reproducible workflow. 

We developed a computational workflow that combines complementary sequence design strategies, including codon optimisation, RNA secondary structure stabilisation, and deep-learning generative design, to generate synonymous mRNA variants spanning a range of predicted translational properties, including the codon adaptation index (CAI), RNA structural features, and AI-based predicted translational efficiency, which are valuable to generate hypotheses for experimental validation. RNAs are subsequently synthesised, purified and evaluated using a high-throughput cell-free luciferase translation assay to quantitatively measure protein production and translational activity. 

As a proof-of-concept, we applied this workflow to GFP, generating six functionally synonymous mRNA constructs comprising two designs predicted to exhibit high translational efficiency, two designs predicted to exhibit low translational efficiency, and the superfold GFP and enhanced GFP reference sequences. Each construct was characterised using a range of computational metrics. Experimental testing of these GFP constructs provided a controlled system for assessing how computationally engineered sequence features and predicted metrics transfer to actual experimental mRNA properties. In summary, by coupling in silico predictions with standardised mRNA production and quantitative translation assays, this Design-Build-Test platform enables systematic benchmarking and iterative refinement of computationally engineered mRNAs and provides a pipeline for systematically evaluating computational mRNA design strategies using experimental feedback. Beyond GFP, the framework is readily extensible to therapeutic mRNAs, providing a scalable platform for next-generation RNA therapeutics.