A key strategy for improving results from generative protein models is "inference-time scaling." This involves generating a vast number of potential structures and then using a separate, fine-tuned scoring model to rank them. This search-and-rank process uncovers high-quality solutions the model might otherwise miss.
Related Insights
Unlike LLMs, parameter count is a misleading metric for AI models in structural biology. These models have fewer than a billion parameters but are more computationally expensive to run due to cubic operations that model pairwise interactions, making inference cost the key bottleneck.
Modern protein models use a generative approach (diffusion) instead of regression. Instead of predicting one "correct" structure, they model a distribution of possibilities. This better handles molecular dynamism and avoids averaging between multiple valid states, which is a flaw of regression models.
To break the data bottleneck in AI protein engineering, companies now generate massive synthetic datasets. By creating novel "synthetic epitopes" and measuring their binding, they can produce thousands of validated positive and negative training examples in a single experiment, massively accelerating model development.
Models like AlphaFold don't solve protein folding from physics alone. They heavily rely on co-evolutionary data, where correlated mutations across species provide strong hints about which amino acids are physically close. This dramatically constrains the search space for the final structure.
AlphaFold's success in identifying a key protein for human fertilization (out of 2,000 possibilities) showcases AI's power. It acts as a hypothesis generator, dramatically reducing the search space for expensive and time-consuming real-world experiments.
As biologics evolve into complex multi-specific and hybrid formats, the number of design parameters (valency, linkers, geometry) becomes too vast for experimental testing. AI and computational design are becoming essential not to replace scientists, but to judiciously sample the enormous design space and guide engineering efforts.
Contrary to trends in other AI fields, structural biology problems are not yet dominated by simple, scaled-up transformers. Specialized architectures that bake in physical priors, like equivariance, still yield vastly superior performance, as the domain's complexity requires strong inductive biases.
Current AI for protein engineering relies on small public datasets like the PDB (~10,000 structures), causing models to "hallucinate" or default to known examples. This data bottleneck, orders of magnitude smaller than data used for LLMs, hinders the development of novel therapeutics.
AI's key advantage isn't superior intelligence but the ability to brute-force enumerate and then rapidly filter a vast number of hypotheses against existing literature and data. This systematic, high-volume approach uncovers novel insights that intuition-driven human processes might miss.
Generative AI alone designs proteins that look correct on paper but often fail in the lab. DenovAI adds a physics layer to simulate molecular dynamics—the "jiggling and wiggling"—which weeds out false positives by modeling how proteins actually interact in the real world.
