Track genomics, neuroscience and molecular biology.
Search papers, labs, and topics across Lattice.
To evaluate computational protein modification under realistic wet-lab constraints, the authors benchmarked six protein language models (PLMs), six LLMs, and five LLM agents across four experimental settings in PFArena. Systematically evaluating these distinct model families is critical for understanding where general-purpose in-context reasoning can substitute for or augment domain-specific biophysical priors. The benchmark reveals that PLMs excel at open-ended single-mutant generation, whereas LLMs and agents decisively outperform them at multi-mutant ranking when target-specific fitness context is available, though all paradigms degrade severely with increased mutation depth.
Domain-specific PLMs dominate zero-shot single-mutation design, but general LLMs and agents overtake them in multi-mutant ranking as soon as target-specific experimental fitness data is provided.
Protein modification requires navigating an immense sequence space, yet wet-lab validation remains low-throughput and costly. Although computational paradigms including protein language models (PLMs), large language models (LLMs), and LLM-based agents have shown promise in protein modification, their relative efficacy across realistic experimental decision-making settings remains unclear. To bridge this gap, we introduce PFArena, a benchmark comprising four controlled task interfaces that cover single-mutant generation and multi-mutant ranking. By providing varying levels of mutation fitness data, PFArena reflects four representative research scenarios characterized by differing degrees of prior experimental context. We assess six PLMs, six LLMs, and five LLM-based agents using complementary metrics to measure both peak and overall protein modification performance. Our evaluation reveals that model performance shifts systematically with the availability of target-specific experimental evidence: PLMs demonstrate proficiency in open-ended single-mutant generation by leveraging protein-specific priors, whereas LLMs and agents perform strongly in multi-mutant ranking, particularly when target-specific fitness data are available. Nevertheless, all model families face fundamental challenges with increasing search-space size and mutation depth. We release our code and benchmark suite to facilitate reproducible research in model-assisted protein modification.