This study systematically assessed the biological knowledge encoded in publicly available LLMs for structured phenotype assignment of microbial species, including state-of-the-art models such as Claude Sonnet 4 and the GPT-5 family of models.
Abstract
Large language models (LLMs) are increasingly used to extract knowledge from text, yet their coverage and reliability in biology remain unclear. Microbial phenotypes are especially important to assess, as comprehensive data remain sparse except for well-studied organisms and they underpin our understanding of microbial characteristics, functional roles, and applications. Here, we systematically assessed the biological knowledge encoded in publicly available LLMs for structured phenotype assignment of microbial species. We evaluated the performance of up to 57 LLMs across different experiments, including state-of-the-art models such as Claude Sonnet 4 and the GPT-5 family of models. Across phenotypes, LLMs reached accurate assignments for many species, but performance varied widely by model and trait, and no single model dominated. Model self-reported confidence is informative, with higher confidence aligning with higher accuracy, and can be used to prioritize phenotype assignment, effectively distinguishing between high- and low-confidence inferences. Overall, our study outlines the utility and limitations of text-based LLMs for phenotype characterization in microbiology.
It is found that many GigaRef singletons belong to a cluster under alternative parameter settings, suggesting that genomic and metagenomic datasets may require dataset-specific clustering configurations, and it is shown that singletons share mutual information with clustered sequences, making them learnable by PLMs and useful for training.
R. Vinod, Samir Char, Ava A. Amini et al.· bioRxiv· 0 citations
Across the included studies, stronger evidence for gLM utility was generally associated with biologically informed or task-aligned model design, including evolutionary alignments, motif-aware objectives, long-context architectures, RNA structural priors, population-aware representations, and domain-specific pretraining.
Mahinaz A. Mashhour, Manal Abdel Wahed, Mai S. Mabrouk· Biochemical and Biophysical...· 0 citations
Protein language models have been increasingly successful on tasks ranging from fitness prediction to functional design, yet what biological knowledge they acquire and where it is encoded within their internal representations remain underexplored. Through a high-resolution layer-by-layer interpretability analysis of 8 models from the ESM2 and AMPLIFY families on 22 concepts from human proteome annotations, we found that these models encode concepts of increasing levels of complexity along their depth: basic physicochemical properties and linear motifs are best captured by early-layer embeddings, secondary structure from subsequent layers, and domain-level semantics from middle layers. Principal component projections of these embeddings showed that they separate biologically meaningful protein groupings, and molecular-biology-inspired interventions demonstrated that pLM embeddings can discriminate phosphomimic-active from inactive mutants. Perhaps surprisingly, we observed that pretraining data and compute had a greater impact on the linear emergence of biological concepts than scaling up parameters. By revealing where biological knowledge is captured in pLMs and which choices shape its emergence, our work offers insights to develop more robust, biologically grounded protein language models.
Shawn T. Whitfield, Tom Marty, Robert M. Vernon et al.· bioRxiv· 0 citations
An embedding-based statistical framework is developed that transforms gene and function descriptions into vector representations, enabling statistical testing of gene-gene and gene-function relationships and quantitative prioritization of de novo functional hypotheses inferred by LLMs.
General-purpose frontier language models are being increasingly utilized for protein-design work, yet their ability to understand and evaluate variant effects remains unclear. Here, we introduce PG-LLM, a benchmark comprising 276 protein-variant prioritization tasks: 217 from ProteinGym and a temporally held-out set of 59 from recently published studies. Each task follows the same format: a language model is asked to rank a list of variant sequences given only the wild-type protein sequence and an assay description with no access to tools, multiple-sequence alignments, or protein structures. We evaluate thirteen language models and 95 published protein predictors on the same variants with the same evaluation metric. Claude Opus 5 (Max) and GPT 5.6 Sol (Max) are the best performing LLMs with Spearman correlations of ρ = 0.406 and 0.402 respectively. Opus 5 outperforms 49 of 95 published protein predictors, including 41 of 46 sequence-only methods, and approaches ESM2-650M at ρ = 0.411, but remains below the leading predictor VenusREM at ρ = 0.523. We observe that variant-ranking performance scales with test-time compute across GPT, Claude, and Gemini models, but gains taper before closing the gap to specialist protein predictors. To address contamination risk, we create a held-out evaluation set with 59 DMS assays from 19 studies whose scores first became public after January 2026. On this set, we observe performance and test time compute scaling trends similar to those on the 217 tasks derived from ProteinGym. PG-LLM shows that tool-free language models capture substantial protein-variant signal, outperforming many sequence-based predictors while remaining below the strongest specialized models.
Rohit Arora, L. Chen, Melissa Du et al.· bioRxiv· 0 citations
GenesetGPT is proposed, an efficient, LLM-based framework that emphasizes both curated biological context and iterative prompt construction, thus enabling realistic summarization of heterogeneous gene sets at scale.
Jack R. Leary, Samantha Pattey, Rhonda L. Bacher· bioRxiv· 0 citations