Can Generative AI Crack the Code of mRNA Design?

New preclinical data suggest generative AI could streamline mRNA design and reduce costly trial-and-error for future therapies

7 Jan 2026

Efforts to make mRNA medicines faster and more predictable are entering a new phase as developers begin to test generative artificial intelligence in the design of drug sequences.

Attention has focused on Raina Biosciences, which in late 2025 released preclinical data on a generative AI platform designed to create mRNA sequences. In selected experimental settings, the approach performed better than conventional design methods, prompting interest among researchers and analysts watching the digital shift in drug development. The company has not announced partnerships linked to the work, but the data circulated widely within the mRNA research community.

The appeal lies in efficiency. Traditional mRNA design depends on repeated cycles of laboratory testing and adjustment. While reliable, the process can be slow and costly. Generative AI aims to change the starting point by producing candidate sequences digitally, guided by targets such as protein expression or stability. Scientists can then move into laboratory testing with a narrower and more informed set of options.

“This is about moving toward intent-driven design, at least early on,” said one industry analyst who follows mRNA platforms. “We have seen similar digital tools reshape other areas of drug discovery, but mRNA is still early in that transition.”

The timing coincides with the broadening of mRNA research beyond pandemic era vaccines into oncology, rare diseases and personalised therapies. Established groups such as Moderna have already invested heavily in digital infrastructure, and analysts say generative AI could complement those systems. Uptake, however, remains uneven as companies weigh the cost of building expertise against uncertain returns.

Constraints remain significant. Generative models rely on large volumes of high quality biological data, which are often fragmented and not consistently shared across the industry. Regulators are also beginning to consider how to assess AI assisted design, particularly when algorithmic choices are less transparent than traditional methods.

For now, generative AI in mRNA design is better seen as an emerging capability than an established standard. Platforms such as Raina’s have moved the idea from theory into early practice, but its long term role in mRNA drug development will depend on whether results continue to hold as more data becomes available.