Gibson SOLA Platform

Drug development is notoriously complex, expensive, and high-risk. One of the biggest determinants of success is the quality of decisions made early in the R&D pipeline—particularly in selecting a promising starting point, or lead candidate.

A growing number of pharmaceutical teams are recognizing that to improve those odds, you need two things: a deep understanding of how biological sequences are constructed, and the ability to combine that knowledge with modern AI tools. The convergence of these two capabilities is reshaping how—and how fast—new therapeutics get developed.

The Importance of Sequence Construction

At its core, every biologic drug originates from a DNA or RNA sequence. These sequences encode proteins, regulate gene expression, or serve as the drug itself (as in mRNA therapeutics). Being able to design and synthesize these sequences accurately is a foundational part of the development process.

Understanding not just what the sequence is, but how it’s built, provides crucial context for predicting biological function, selecting targets, optimizing efficacy, and avoiding failure-prone candidates.

Legacy Limitations: The Bottlenecks of Traditional DNA Synthesis

Historically, DNA and RNA synthesis has relied heavily on external vendors. Researchers submit sequence designs, then wait—often for days or weeks—for a shipment. The process is slow, subject to delays, and introduces risk at the most critical phase: early validation.

This bottleneck reduces the pace of experimentation and lengthens the feedback loop for decision-making. Worse, if the sequence has synthesis errors, it might generate false results or require rework.

Accelerating R&D with Gibson SOLA

Telesis Bio’s Gibson SOLA Platform addresses this head-on. It enables researchers to synthesize high-fidelity DNA or mRNA in-house, overnight.

Rather than constructing long DNA sequences one nucleotide at a time using traditional phosphoramidite chemistry, Gibson SOLA uses enzymatic methods to assemble short, pre-designed oligo-nucleotides from a universal stock library. This approach significantly lowers error rates and speeds up the process (see figure 1).

The result: researchers gain control over the build process, reduce dependency on third parties, and significantly compress development timelines.

The AI Factor: Data-Driven Discovery

AI models in drug discovery span a range of machine learning approaches, from traditional algorithms like random forests and support vector machines to advanced deep learning frameworks such as convolutional neural networks (CNNs) and transformer-based architectures. Each model type is suited to different data modalities—CNNs might analyze 3D protein structures, while transformers excel at interpreting long DNA or RNA sequences. These models are trained on vast datasets that include genomic annotations, high-throughput screening data, molecular docking simulations, and clinical trial outcomes. Once trained, they can predict which compounds are likely to bind a biological target, how a mutation might affect protein folding, or whether a molecule is likely to pass safety thresholds. Importantly, they also generate probabilistic confidence scores, helping scientists prioritize candidates more objectively. This computational layer not only speeds up decision-making but can uncover previously hidden biological insights that push discovery into entirely new territory.

However, the quality of an AI’s output is tightly coupled to the quality of its input. Inaccurate or low-fidelity sequence data undermines predictive performance. Worse, gaps in data caused by external service providers who are unwilling to build the desired sequences result in missed connections and insights. That’s why pairing high-quality, real-time synthesis tools like Gibson SOLA with AI workflows is such a powerful combination (figure 3).

When AI Meets Fast, Accurate Synthesis

AI models can identify promising lead candidates based on predictive models. As more data is fed into the models, they improve in accuracy of predicted target candidates and mechanisms. The Gibson SOLA Platform allows those candidates to be synthesized on-demand for accelerated insights. Together, they create a tight feedback loop (figure 3). This cycle, repeated rapidly, allows teams to test more leads, refine ideas faster, and avoid dead-ends early in the process.

Empowered R&D Teams

Bringing synthesis in-house and combining it with AI models accelerates timelines and also changes how teams think and operate.

• Agility: Scientists can test hypotheses the same week they’re formed.
• Scalability: The system supports everything from pilot studies to production.
• Innovation: With fewer logistical delays and more control, teams can take more creative risks.

The result is a pipeline that moves faster, costs less, and is more likely to produce viable therapeutics.

Conclusion

Pharma R&D teams need to make smarter bets earlier in the pipeline. That means starting with better information, better tools, and better feedback loops. Platforms like Gibson SOLA give researchers control over one of the most foundational aspects of drug development: sequence construction. The cyclic process of build, test, iterate happens more quickly with potentially more valuable insights obtained in each cycle.

In an industry where every delay can cost millions, and every failed candidate sets back progress, that kind of speed and precision isn’t just convenient, it’s transformative.