Biology subjects

Uehara, M.

Publications and source records attributed to Uehara, M..

8 recordsLinked to original sources

bioRxiv2026

LifeSciBench: Evaluating Language Models on Realistic, Expert-Level Tasks in the Life Sciences

We introduce LifeSciBench, a benchmark of 750 expert-authored tasks designed to evaluate whether language models can handle realistic life science research work. The majority of existing life sciences benchmarks have a narrow scope or are purely knowledge-based, and therefore fail to capture the complexity of real-world research, which often involves ambiguities and requires the accurate execution of multiple dependent judgment calls. Additionally, almost all existing benchmarks span at best a small collection of subdomains within the life sciences; there is at present no existing life sciences benchmark with both the requisite breadth and depth required to convincingly measure proficiency in real-world professional research settings. LifeSciBench addresses this gap by spanning seven representative scientific workflows and seven life science domains, with each constituent task paired with a human expert-written rubric. Across five frontier and domain-specialized models, GPT-Rosalind performs best, with a task-weighted mean normalized rubric score of 0.576 and a task-weighted response pass rate of 36.1% (response-level values are first averaged within each task, and the resulting task-level values are then averaged with equal weight). LifeSciBench remains unsaturated, with 171 tasks (22.8%) having no observed passing response from any evaluated model and 261 tasks (34.8%) having a best-model pass rate below 20%. LifeSciBench therefore serves as a high-resolution evaluation of practical scientific reasoning and operational decision-making in the life sciences.

bioinformatics↗

bioRxiv2026

Language Modeling Materializes a World Model of Protein Biology

Proteins are fundamental to life. The full extent of their biology is beyond our ability to characterize with experimental approaches in the physical laboratory. Accurate digital representations could accelerate the discovery of protein biology through virtual experiments. We propose language modeling to learn unified and general representations that can be scaled to all of protein biology. Building on these representations, we develop a structure prediction model that exceeds the performance of established methods for biomolecular complex prediction across benchmarks, including for the interactions of antibodies with their targets. A simple search procedure yields high experimental success rates for the discovery of proteins with nanomolar binding affinities for both miniproteins and single-chain antibodies, a modality critical for therapeutic design. Study of the concepts in the language models representation space reveals a systematic organization aligned with the reductionist understanding of proteins developed through empirical science. Leveraging this organization, we generate a comprehensive map of protein biology encompassing over 6.8 billion sequences and 1.1 billion predicted structures, identifying connections across known and unknown biology. As a whole, this shows language modeling as a powerful substrate for representing the biology of proteins, operating across scales from the prediction and design of protein interactions at the atomic level, to identifying properties of proteins at different levels of granularity and abstraction, to the scale of mapping connections between proteins across billions of years of evolution.

bioinformatics↗

bioRxiv2025

Generative AI-driven artificial DNA design for enhancing inter-species gene activation and enzymatic degradation of PET

Conventional approaches to heterologous gene expression rely on codon optimization, which is limited to swapping synonymous codons and often fails to capture deeper adaptive changes. In contrast, naturally evolved orthologous genes between species often differ by more than just synonymous substitutions - they can include non-synonymous mutations, insertions, and deletions that confer functional adaptation to different host contexts. Here we present OrthologTransformer, a Transformer-based deep learning model that converts orthologous genes between species by learning from large-scale orthologous gene datasets curated for high-quality sequence alignments. The model recapitulates the full spectrum of evolutionary differences - from synonymous codon swaps to amino acid-changing mutations and indels - to predict a coding sequence optimized for a target species while preserving the proteins function. In extensive validation across diverse bacterial species pairs, OrthologTransformer significantly increased the conversion accuracy of generated genes to native target sequences compared to the original source genes, even for pairs with stark disparities in GC content and optimal growth temperature. The Transformers context-aware designs also favored conservative amino acid substitutions, maintaining protein functional integrity. As a proof-of-concept, an OrthologTransformer-designed PETase gene for Bacillus subtilis from the host species Ideonella sakaiensis was synthesized and expressed, yielding robust PET plastic-degradation activity that surpassed synonymous codon-optimized controls. These results establish OrthologTransformer as a powerful tool for de novo cross-species gene adaptation, transcending the limits of traditional codon optimization and enabling more effective heterologous gene performance in synthetic biology applications.

bioinformatics↗

bioRxiv2025

SpatialAgent: An Autonomous AI Agent for Spatial Biology

Advances in AI are transforming scientific discovery, yet spatial biology, a field that deciphers the molecular organization within tissues, remains constrained by labor-intensive workflows. Here, we present SpatialAgent, an autonomous AI agent for spatial biology research. SpatialAgent couples large language models with a Plan-Act-Conclude architecture, dynamic tool and skill retrieval, multimodal interpretation, and verification modules that audit generated claims. It supports the full discovery loop, from gene-panel design and multimodal annotation to trajectory inference, cell-cell communication analysis, imputation, and hypothesis generation. Across human and mouse brain, heart, tonsil, colon, and prostate datasets, SpatialAgent outperformed established computational baselines and matched or surpassed expert scientists in key tasks. In open-ended case studies, it recovered known tissue organization and generated spatially grounded hypotheses. In a prospective mouse prostate cancer Xenium study, it designed a compact 100-gene add-on panel that profiled 4.2 million cells across 21 samples, improved cell-type and malignant-state prediction, and captured spatially structured tumor and microenvironment programs. SpatialAgent establishes a framework for autonomous and collaborative discovery in spatial biology.

bioinformatics↗

bioRxiv2023

Decoding host-microbiome interactions through co-expression network analysis within the non-human primate intestine

The gut microbiome affects the health status of the host through complex interactions with the hosts intestinal wall. These host-microbiome interactions may spatially vary along the physical and chemical environment of the intestine, but these changes remain unknown. This study investigated these intricate relationships through a gene co-expression network analysis based on dual transcriptome profiling of different intestinal sites--cecum, transverse colon, and rectum-- of the primate common marmoset. We proposed a gene module extraction algorithm based on graph theory to find tightly interacting gene modules of the host and the microbiome from a vast co-expression network. The 27 gene modules identified by this method, which include both host and microbiome genes, not only produced results consistent with previous studies regarding the host-microbiome relationships, but also provided new insights into microbiome genes acting as potential mediators in host-microbiome interplays. Specifically, we discovered associations between the host gene FBP1, a cancer marker, and polysaccharide degradation-related genes (pfkA and fucI) coded by Bacteroides vulgatus, as well as relationships between host B-cell specific genes (CD19, CD22, CD79B, PTPN6) and a tryptophan synthesis gene (trpB) coded by Parabacteroides distasonis. Furthermore, our proposed module extraction algorithm surpassed existing approaches by successfully defining more functionally related gene modules, providing insights for understanding the complex relationship between the host and the microbiome. IMPORTANCEWe unveiled the intricate dynamics of the host-microbiome interactions along the colon by identifying closely interacting gene modules from a vast gene co-expression network, constructed based on simultaneous profiling of both host and microbiome transcriptomes. Our proposed gene module extraction algorithm, designed to interpret inter-species interactions, enabled the identification of functionally related gene modules encompassing both host and microbiome genes, which was challenging with conventional modularity maximization algorithms. Through these identified gene modules, we discerned previously unrecognized bacterial genes that potentially mediate in known relationships between host genes and specific bacterial species. Our findings underscore the spatial variations in host-microbiome interactions along the colon, rather than displaying a uniform pattern throughout the colon.

microbiology↗

bioRxiv2023

Foxp- and Skor-family proteins control differentiation of Purkinje cells from Ptf1a and Neurogenin1-expressing progenitors in zebrafish

Cerebellar neurons such as Purkinje cells (PCs) and granule cells (GCs) are differentiated from neural progenitors expressing proneural genes. Zebrafish mutants of proneural genes ptf1a and neurogenin1 showed a reduction or loss of PCs, GABAergic interneurons (INs), and reduced expression of GC progenitor genes atoh1a/b/c. Lineage tracing revealed that the ptf1a-expressing progenitors gave rise to PCs, INs, and a part of GCs in zebrafish. These data indicate that the ptf1a/neurognin1-expressing neural progenitors can generate a variety of cerebellar neurons. In this study, we found that genes encoding transcriptional regulators Foxp1b and Foxp4, as well as Skor1b and Skor2, which are reportedly expressed in PCs, were not expressed in ptf1a;neurogenin1 mutants. foxp1b;foxp4 mutants showed a strong reduction in PCs, while skor1b;skor2 mutants completely lacked PCs but instead displayed an increase in immature GCs. Misexpression of skor2 in GC progenitors expressing atoh1c suppressed GC fate. These data indicate that Foxp1b/4 and Skor1b/2 function as key transcriptional regulators in the initial step of PC differentiation from ptf1a/neurogenin1-expressing neural progenitors, while Skor1b and Skor2 control PC differentiation by suppressing their differentiation into GCs.

developmental biology↗

bioRxiv2023

Mass cytometry identifies characteristic immune cell subsets in bronchoalveolar lavage fluid from interstitial lung diseases

Immune cells have been implicated in interstitial lung diseases (ILDs), although their phenotypes and effector mechanisms remain poorly understood. To better understand these cells, we conducted an exploratory mass cytometry analysis of immune cell subsets in bronchoalveolar lavage fluid (BALF) from patients with idiopathic pulmonary fibrosis (IPF), connective-tissue disease (CTD)-related ILD, and sarcoidosis, using two panels including 64 markers. Among myeloid cells, we observed the expansion of CD14+ CD36hi CD84hi monocyte populations in IPF. These CD14+ CD36hi CD84hi subsets were also increased in ILDs with a progressive phenotype, particularly in a case of acute exacerbation (AEx) of IPF. Analysis of B cells revealed the presence of cells at various stages of differentiation in BALF, with a higher percentage of IgG memory B cells in CTD-ILDs and a trend toward more FCRL5+ B cells. These FCRL5+B cells were also present in the patient with AEx-IPF and sarcoidosis with advanced lung lesions. Among T cells, we found increased levels of IL-2R+TIGIT+ LAG3+ CD4+T cells in IPF, increased levels of CXCR3+ CD226+ CD4+ T cells in sarcoidosis, and increased levels of PD1+ TIGIT+ CD57+ CD8+ T cells in CTD-ILDs. Together, these findings underscore the diverse immunopathogenesis of ILDs.

immunology↗

bioRxiv2021

Development of a metatranscriptomic analysis method for multiple intestinal sites and its application to the common marmoset intestine

BackgroundThe intestinal microbiome is closely related to host health, and metatranscriptomic analysis can assess the functional activity of microbiomes by quantifying the bacterial gene expression level, which helps to elucidate the interaction between the microbiome and the environment. However, functional changes in the microbiome along the host intestinal tract remain unknown, and previous analytical methods have limitations, such as potentially overlooking unknown genes due to dependence on existing databases and being unable to take full advantage of metatranscriptome to reveal the functional change among multiple environments. ResultTo close these gaps, we develop a novel method that integrates metagenome and metatranscriptome to analyze the functional activity of microbiomes between intestinal sites. This method reconstructs a reference metagenomic sequence across multiple intestinal sites, allowing the gene expression levels of microbiome including unknown bacterial genes to be compared between multiple sites. As a result of applying this method to metatranscriptomic analysis in the intestinal tract of common marmoset, the reconstructed metagenome covered most of the expressed genes and it revealed that the changes in bacterial gene expressions among the caecum, transverse colon, and faeces were more dynamic and sensitive to environmental shifts than its abundance. In typical, the coenzyme synthesis gene and antibacterial resistance gene were more highly expressed in the caecum and transverse colon than in faeces, while there was no significant change in abundance of these genes. ConclusionOur findings demonstrate that an analytical method that integrates metagenome and metatranscriptome in multiple intestinal sites captures functional changes in the microbiomes at the gene resolution level.

microbiology↗