Biology subjects

Xiong, Y.

Publications and source records attributed to Xiong, Y..

At least 19 recordsLinked to original sources

bioRxiv2026

GLOM TGITO: multiscale protein modelling and multiview landscape-guided sequence discovery

Motivation: Offline protein design must seek high-activity variants while retaining support for their predicted performance. The challenge is to explore multiple promising regions without confining the search to average-performing sequences. We test whether multiscale representations and local evidence can guide this search. Results: GLOM-TGITO links a global-local fluorescence predictor to multiview landscape-guided optimization. Across 31,434 measured ppluGFP2 sequences, GLOM achieved random/depth-holdout RMSEs of 0.170/0.247 versus 0.247/0.324 for the strongest SaProt-650M controls. Averaged across three training seeds, 96.3%/97.3% of prediction-ranked Top 100 test variants were measured bright (log10 fluorescence >=4.2). Search from 900 peak regions yielded 2,594 distinct candidate sequences. The selected Top 100 showed 2.9-8.2% lower mean representation deviation than four comparator pipelines; 38 versus 11-16 met joint novelty-reliability filters. Deviation identified measured test errors >0.1 (AUROC 0.773). The first molecular-dynamics run of the selected Top 30 identified smaller surface hydrophobic patches in 20 candidates relative to their starts, with representative variants showing more persistent chromophore contacts or narrower torsional fluctuations.

bioinformatics↗

bioRxiv2026

NeCS-Flow: Necrosis-Guided Contrast-Enhanced MRI Synthesis with Rectified-Flow Refinement

Background: Gadolinium-based contrast agents raise concerns about tissue deposition, adverse reactions, and increased costs. Purpose: To develop and validate a deep learning pipeline for gadolinium-free post-contrast T1w brain MRI synthesis incorporating necrotic scaling and Posterior-Mean Rectified Flow (PMRF) refinement, and to evaluate diagnostic utility through a blinded multi-reader study. Study Type: Retrospective. Population: 454 patients with adult-type gliomas (grade 2-4) from four publicly available datasets. Training used 374 patients; internal testing used 60 held-out patients; external validation used 20 patients. Field Strength/Sequence: 1.5-3.0 T; pre-contrast T1w, T2w, and FLAIR as inputs; post-contrast T1w as synthesis target. Assessment: A three-stage pipeline comprising nnU-Net Necrotic core segmentation, necrotic-scaling-guided U-Net Contrast Synthesis, and PMRF perceptual refinement (NeCS-Flow) was evaluated using whole-brain and tumor-region similarity metrics, enhancing tumor Dice scores, perceptual quality metrics, and a blinded three-reader study assessing diagnostic accuracy, enhancement detection, image quality, and real-versus-synthetic discriminability. Statistical Tests: Paired t-tests and two one-sided tests for model comparison; Fisher's exact test and two-sample t-tests for reader study comparisons, with p < 0.05 considered significant. Results: Necrotic scaling consistently improved real versus synthetic enhancing tumor Dice internally (HGG: 0.793 to 0.816) and externally (HGG: 0.465 to 0.612). NeCS-Flow outperformed two benchmark architectures on all quantitative metrics in internal and external testing, although absolute enhancing tumor accuracy declined for all models on the external cohort. PMRF refinement improved perceptual realism across both tumor types, reducing both FID and KID. Readers distinguished synthetic from real images only modestly above chance (overall accuracy 61.7%), and no significant differences between real and synthetic images were found for any clinical metric. Data Conclusion: NeCS-Flow synthesizes perceptually realistic and diagnostically equivalent post-contrast T1w brain MRI, with necrotic scaling improving enhancing tumor delineation and PMRF refinement improving perceptual realism across gliomas. Performance on an external cohort was preserved relative to benchmark methods, though reduced in absolute terms.

bioinformatics↗

bioRxiv2026

Expandable DNA-origami rings as a nanomechanical platform for studying disordered nucleoporins

Phe-Gly-rich nucleoporins (FG-Nups) within the nuclear pore complex (NPC) form a diffusion barrier that controls the exchange of macromolecules between the nucleus and cytoplasm. The behaviors of these intrinsically disordered FG-Nups are fundamental to the NPC function but remain difficult to probe within the dynamic, native nuclear pore. To bridge this knowledge gap, here we build expandable DNA-origami rings with tunable elasticity and study FG-Nups site-specifically tethered in such mechanically responsive NPC mimics. The elastic deformation of DNA rings, measured by single-particle TEM, reveals the type (attraction or repulsion) and strength of FG-Nup interactions that shape the collective morphology of these unstructured proteins. Analyzing homotypic interactions of five FG-Nup species finds diverse FG-Nup behaviors that can be classified as repulsive/extended (Nsp1), moderately cohesive/extended (Nup62, Nup153, and Nup214), and strongly cohesive/compact (Nup98). Furthermore, we examine how FG-Nups' primary sequence, O-glycosylation, grafting density, and binding to nuclear transport receptors (NTRs) impact their biophysical state. For example, O-GlcNAcylation reduced FG-Nup cohesiveness, switching Nup98 from a collapsed state to more extended conformations. Our work thus elucidates spatially confined Nup-Nup and Nup-NTR interactions within an NPC-like nanopore, suggests a previously underappreciated role of FG-Nups in modulating NPC dilation, and establishes a generalizable method for studying multivalent interactions among disordered proteins.

biophysics↗

bioRxiv2026

Controlling membrane tension by dynamic DNA nanorings

Membrane tension, defined as the energy cost to expand the membrane surface area, is a fundamental mechanical property of the lipid bilayer and an important modulator of various cellular processes. Liposomes provide model systems to study protein-mediated membrane dynamics, where membrane tension can be controlled by osmotic pressure and micropipette aspiration. However, it is challenging to individually control membrane tension of nanometer-sized liposomes, such as small unilamellar vesicles (SUVs), that are widely used for modeling subcellular membrane structures and drug delivery. To bridge this technological gap, we present reconfigurable DNA origami rings for dynamically controlling membrane tension of sub-100 nm liposomes. The DNA nanorings template the formation of uniformly sized SUVs and undergo trigger-responsive dilation and contraction, thus applying controlled and reversible mechanical stress to the SUV membranes. Electron microscopy analyses show deformed liposomes with ~2.5% expanded membrane area following ring dilation. The corresponding increase in membrane tension opens a mechanosensitive ion channel MscL with gating tension of ~10 mN/m. Selectively manipulating a mixture of SUV populations enables timed release of their distinct molecular cargos, either concurrently or one at a time. Furthermore, vesicles with expanded membranes are more conducive to SNARE-mediated fusion. The nanomechanical device thus provides a programmable platform for engineering the mechanics of synthetic nano-vesicles and studying membrane tension modulated processes at the molecular level.

biophysics↗

bioRxiv2026

dubTAGs enable on-demand stabilization for tunable and reversible control of endogenous protein levels

Precise and rapid control over cellular protein levels is essential to dissect complex biological systems. Chemical genetic approaches such as dTAG, in which a target is fused to a degron tag (FKBP12F36V) and degraded upon small molecule-mediated recruitment of E3 ligases, have enabled rapid and tunable control over protein abundance. However, no analogous tool exists to precisely increase protein levels and actively reverse dTAG-mediated degradation. Here, we developed heterobifunctional small molecules (dubTAGs) that stabilize FKBP12F36V-tagged proteins by recruiting endogenous deubiquitinases. Utilizing stem cell-derived cranial neural crest cells (CNCCs) in which the transcription factors SOX9 or TWIST1 are endogenously tagged with FKBP12F36V, we identified OTUB1- or USP7-recruiting heterobifunctional molecules that demonstrated effective target stabilization and ternary complex formation. We demonstrate that dubTAG-mediated protein stabilization is dependent on deubiquitinase recruitment, target-specific, and can tunably and rapidly reverse dTAG-mediated degradation. We applied dubTAGs to assess how stabilizing endogenous SOX9 impacts chromatin accessibility in CNCCs, finding both monotonic and non-monotonic regulatory element responses that are driven by distinct sequence features. dubTAGs are readily applicable tools for investigating the effects of elevated protein levels and tunably reversing targeted degradation, enabling new approaches to study protein dosage effects in development, disease, and therapeutic discovery.

molecular biology↗

bioRxiv2026

Sex differences in pediatric EEG inter-subject correlation during naturalistic movie watching: a large-scale characterization of the Healthy Brain Network EEG dataset

Inter-subject correlation (ISC) of EEG during naturalistic viewing is increasingly used as a candidate pediatric biomarker, yet its behavior at cohort scale remains uncharacterized. We characterize ISC in 1143 children and adolescents (765 male, 378 female, ages 5 to 21) from the Healthy Brain Network EEG dataset viewing four naturalistic film clips. Three results follow. First, ISC separates narrative from abstract stimuli by 1.7 to 2.4 fold and exceeds a resting-state baseline by 42 to 99 fold, and the previously reported developmental decline replicates throughout. Second, ISC is sex-dependent: males exceed females in every movie (t = 9.7 to 13.5; Cohen's d = 0.61 to 0.85), concentrated in delta and theta, peaking at a frontocentral cluster and at ages 11 to 14. The direction matches an earlier report where the effect was marginal, and we resolve structure that sample could not. It survives three demographic and clinical robustness analyses, same-sex and size-matched templates, ocular component removal, and exclusion of the releases overlapping that report, and is attenuated, by approximately 10 percent adjusting for interpolation count and by 23 percent in recordings free of cluster interpolation, but not explained by differential channel interpolation. Evoked response magnitude, previously proposed as an alternative account, does differ by sex here (d = 0.55 to 0.60) but mediates only 8.6 to 18.1 percent of the effect. Third, ISC predicts none of four CBCL bifactor dimensions after FDR correction at any feature resolution: zero of 16 grand-mean, 2064 per-channel and 16 cluster tests, with confidence intervals excluding standardized associations beyond |{beta}| = 0.106.

neuroscience↗

bioRxiv2026

Integrating Narrow-Window DIA with AI-Powered Search Enables Deep and Reliable Functional Proteomics

Narrow-window data-independent acquisition (nDIA) is emerging as a powerful technique for bottom-up proteomics. Here, we systematically benchmarked nDIA, wide-window DIA (wDIA), narrow-window data-dependent acquisition (nDDA), and wide-window DDA (wDDA) for rapid, single-shot proteomic analysis. For data processing, we introduced Tesorai Search, a new search engine leveraging a large pre-trained model and compared it with DIA-NN and FragPipe across both DIA and DDA datasets. Among 12 acquisition-analysis pipelines evaluated, nDIA combined with DIA-NN and Tesorai Search delivered the highest proteome coverage, identifying 10,255 and 10,766 protein groups from benchmark samples, respectively. Both search engines maintained rigorous false-discovery rate (FDR) control. While nDIA generally outperformed nDDA in sensitivity, FragPipe-DDA+ approach proved to be the most sensitive within the nDDA pipelines. However, entrapment analyses indicate that this sensitivity comes at the cost of less robust FDR control compared to Tesorai Search. As a proof of concept, we applied nDIA-MS to 17 cancer cell lines harboring DNA damage response (DDR) gene knockouts, successfully detecting significant downregulation of all targeted proteins and uncovering 81 DDR-related proteins modulated in at least one cell line. These results underscore nDIA-MS, together with DIA-NN and Tesorai Search, as a robust and scalable platform for high-throughput functional proteomic screening.

systems biology↗

bioRxiv2026

Expanding the Ligandable Chemical Space of OTUB1 through Discovery of a Four-Membered-Ring Recruiter Chemotype

Deubiquitinase-targeting chimeras (DUBTACs) have emerged as a promising strategy for targeted protein stabilization, but their broader application remains limited by the scarcity of ligandable deubiquitinase recruiters. Here, we report a previously unexplored four-membered-ring OTUB1 recruiter chemotype. Through systematic structure-activity relationship studies, we identified compound 21 (MS2159) as a potent and selective covalent OTUB1 ligand. Biochemical and intact protein mass spectrometric analyses demonstrated that MS2159 selectively engages the non-catalytic C23 residue of OTUB1, shows minimal reactivity toward other tested proteins, and preserves OTUB1 deubiquitinase activity. Conjugation of MS2159 with the CFTR ligand lumacaftor yielded compound 25 (MS2134), which effectively stabilized {Delta}F508-CFTR. Collectively, these findings establish a new OTUB1 recruiter scaffold, expand the ligandable chemical space of OTUB1, and provide additional opportunities for developing next-generation DUBTACs.

biochemistry↗

bioRxiv2026

A small RNA guides a post-transcriptional regulatory protein to its target

RNA-binding proteins are central to post-transcriptional regulation and are generally thought to recognize their target RNAs through direct binding. Here, we uncover a previously unrecognized mechanism in which a small RNA (sRNA) guides a protein regulator to specific target mRNAs, enabling programmable post-transcriptional control. By mapping in pathogenic Escherichia coli the RNA interactome of CsrA, a global post-transcriptional regulator, we identify ~800 ternary complexes in which CsrA simultaneously binds an sRNA and an mRNA. Focusing on one class of these complexes, we show that PasE, a newly discovered sRNA, directs CsrA to virulence-associated mRNAs that are otherwise not targeted by CsrA, resulting in their repression. Altering the PasE seed sequence redirects CsrA to selected mRNAs, establishing a modular RNA-guided platform for gene regulation. Together, our findings reveal a new and likely widespread mechanism of bacterial post-transcriptional regulation and provide a framework for a new type of RNA-based synthetic regulation.

microbiology↗

bioRxiv2026

Targeting the NuRD Component, CHD4, Impairs Foxp3+ Treg Cell Production and Function and Promotes Anti-Tumor Immunity

Little is known about why Foxp3 regulatory T (Treg) cells require at least three HDAC1/HDAC2-containing chromatin-remodeling complexes (NuRD, Sin3 and CoREST), or whether selective disruption of these complexes can be exploited to enhance antitumor immunity. Here, we investigated the role of chromodomain helicase DNA- binding protein 4 (CHD4), the ATP-dependent remodeling subunit of the NuRD complex, in Treg biology. Conditional deletion of Chd4 in Foxp3 Tregs resulted in severe systemic autoimmunity and early lethality, accompanied by reduced Foxp3 expression, impaired Treg suppressive function, and loss of Treg lineage stability. Transcriptomic analyses demonstrated that CHD4 deficiency closely phenocopied Hdac2 deletion, whereas quantitative proteomic analyses revealed that CHD4 assembles into highly conserved NuRD complexes in both Treg and conventional CD4 T cells. These findings indicate that the selective dependence of Tregs on CHD4 does not arise from the formation of lineage-specific protein complexes but rather from the unique epigenetic program maintained by CHD4-containing chromatin-remodeling complexes that is required for Treg differentiation and stability. Using a novel cellular target-engagement platform, we identified CH41, a potent small-molecule inhibitor of CHD4 that recapitulated the effects of genetic CHD4 ablation on Treg function. Pharmacological inhibition of CHD4 impaired intratumoral Treg accumulation and function and significantly inhibited the growth of lung and hepatocellular carcinomas in immunocompetent, but not immunodeficient, mice, without inducing systemic autoimmunity. Collectively, our findings identify CHD4 as a critical epigenetic regulator of Treg lineage stability and establish pharmacological targeting of the CHD4/NuRD axis as a promising strategy to selectively disrupt tumor-associated Tregs and enhance antitumor immunity.

cancer biology↗

bioRxiv2026

Brassinosteroid-auxin crosstalk shapes rice flag leaf angle to boost dense planting yield

Flag leaf angle (FLA) critically determines rice yield potential under dense planting conditions. As two pivotal phytohormones determine rice FLA, the antagonistic interaction between brassinosteroid (BR) and auxin remains largely uncharacterized. We here demonstrate that BR signaling reduces auxin biosynthesis to control FLA via regulating the biosynthesis of secondary cell wall (SCW). Genetic evidence demonstrates OsYUC8 mutants disrupt SCW formation, leading to increased FLA. At the molecular level, the BR-related transcription factor OsBZR1 directly binds to and represses OsYUC8 promoter activity, thereby fine-tuning auxin-mediated SCW biosynthesis. Field evaluations reveal that osbzr1 mutants display optimized flag leaf architecture and improved yield performance under high-density cultivation. Our study not only delineates the antagonistic BR-auxin interaction governing FLA but also establishes a genetic strategy for manipulating crop architecture to maximize yield in dense planting conditions. One-sentence summaryOsBZR1-mediated repression of OsYUC8-driven auxin biosynthesis modulates secondary cell wall formation to optimize flag leaf architecture, providing a genetic strategy for maximizing yield under high-density cultivation.

plant biology↗

bioRxiv2026

Coupling Luciferase Activation and BRET Enables High-Contrast Red-Window Calcium Imaging

Genetically encoded bioluminescent calcium indicators enable excitation-free imaging of Ca2+ dynamics with low background, no photobleaching, and improved tissue penetration. However, many exhibit lower dynamic ranges than their fluorescent counterparts, motivating the development of new sensor configurations, particularly for imaging in the red optical window, where tissue absorption and scattering are reduced. Here, we developed HyBRIC and HyBRIC2, two hybrid indicators that integrate Ca2+-dependent luciferase activation and bioluminescence resonance energy transfer (BRET) within a single molecular architecture. When paired with pyCTZ luciferin, they exhibited an approximately 55-fold Ca2+-dependent signal increase in the red optical window in vitro. Compared with HyBRIC, HyBRIC2 showed improved brightness and responsiveness in mammalian cells. It robustly resolved induced Ca2+ dynamics in HEK293T and HeLa cells, with particularly strong red-window responses. Together, these findings expand the limited repertoire of responsive red-window bioluminescent calcium indicators and establish hybrid multiplicative sensing as a modular strategy for improving biosensor contrast.

biochemistry↗

bioRxiv2026

Integrated Patient-Derived Xenograft and Patient-Derived Cell Models Reveal Therapeutic Vulnerabilities Beyond Standard-of-Care Therapy in Endometrial Cancer

Endometrial cancer (EC), the most common gynecologic malignancy in the USA, has seen limited improvement in patient outcomes over recent decades, underscoring the need for relevant preclinical models. To address EC heterogeneity, we established an integrated platform of patient-derived xenografts (PDXs) and matched patient-derived primary cancer cells (PDCs) for disease modeling and systematic drug sensitivity testing. Fresh tumor specimens (n=103) were collected from EC patients to generate PDXs in immunodeficient mice and corresponding PDCs. Fifty-three PDX models were successfully established (52% engraftment rate), with higher success observed in high-grade, recurrent, metastatic tumors (70%), compared with their low-grade counterparts (56%). Histopathologic and immunohistochemical analyses confirmed that PDX tumors faithfully preserved morphology, hormone receptor status, and intertumoral heterogeneity across multiple passages. Using 13 PDC models, we performed an unbiased screening of 179 FDA-approved oncology drugs, revealing marked intertumoral variability in drug response. Almost all PDC models exhibited limited sensitivity to NCCN-recommended therapies, highlighting the need for alternative treatment strategies. In contrast, multiple FDA-approved agents including epigenetic modulators, dual PI3-kinase/HDAC inhibitors, topoisomerase II inhibitors, and proteasome inhibitors demonstrated potent antitumor activity. Importantly, a low-dose combination of the DNA methyltransferase inhibitor 5-azacytidine and the histone deacetylase inhibitor romidepsin significantly suppressed tumor growth across six independent PDX models. Together, these findings establish a comprehensive PDX and PDC platform as a robust translational resource. By capturing the histopathologic and molecular diversity of EC and identifying clinically actionable therapeutic advantages, including an epigenetic combination regimen, this study offers a translational resource for preclinical drug evaluation.

cancer biology↗

bioRxiv2026

CryoGO enables high-resolution structural profiling of endogenous cellular macromolecules

Resolving macromolecular structures within their native cellular environment is essential for connecting molecular architecture to physiological function, yet the field lacks accessible, high-throughput methods for generating suitable specimens from cells. Here, we introduce cryoGO (on-Grid Opening cryo-electron microscopy), a simple, rapid, and scalable strategy that mechanically opens cells directly on EM grids to produce cell-derived specimens for high-resolution single-particle cryo-EM. This approach enables near-atomic structure determination of diverse endogenous macromolecular assemblies spanning a broad molecular-weight range. We show that cryoGO captures the rich ribosomal conformational and compositional landscape while preserving physiological state distributions and aspects of spatial heterogeneity originating from cells. Furthermore, this rapid workflow enables time-resolved structural profiling, capturing both long-term cellular adaptation and acute remodeling on timescales of seconds. Compatible with standard cryo-EM infrastructure and diverse biological samples, while requiring only small numbers of cells, cryoGO lowers the technical barrier to high-resolution native structural biology.

biophysics↗

bioRxiv2026

Combinatorial regulation of cellular rotation by CUL-3-actomyosin-dependent oriented division, eggshell geometry, and Ras–MAPK signaling during dorsal–ventral axis establishment in Caenorhabditis elegans

Cellular rotation is an understudied mechanism that regulates animal morphogenesis. In C. elegans, the dorsal-ventral axis is established when the two-cell-stage AB cell rotates within the eggshell as it divides, generating the diamond-shaped blastomere arrangement at the four-cell stage that enables distinct cell fate specification. Multiple mechanisms, including actomyosin-dependent oriented division, chiral cortical flow, and eggshell shape, have been proposed to regulate this arrangement, but whether these represent conflicting hypotheses or co-acting mechanisms remains unclear. Here, we show that CUL-3-actomyosin-dependent oriented division, eggshell geometry, and the Ras-MAPK signaling pathway regulate distinct steps of cellular rotation. AB cell rotation occurred in two distinct phases: Phase I during AB cytokinesis and Phase II during cytokinesis of the neighboring P1 cell. Quantitative analysis revealed that CUL-3-actomyosin-dependent oriented division is the only one of these three pathways that regulates the AB division axis before anaphase. Actomyosin-dependent oriented division and eggshell geometry were both required for Phase I rotation, whereas Phase II rotation was independent of eggshell geometry. We further identified the Ras-MAPK signaling pathway as a regulator of AB cell rotation that acts independently of eggshell geometry. Strikingly, the CUL-3-actomyosin-dependent pathway may have two distinct roles: first, specifying the AB division axis, and second, correcting the division axis in all cell types during cytokinesis. Together, these functions contribute significantly to cellular rotation and dorsal-ventral axis establishment.

Cell Biology↗

bioRxiv2026

Complete elucidation and heterologous reconstruction of the biosynthetic pathway of camptothecin

Camptothecin derivatives are first-line anticancer drugs used worldwide for the treatment of diverse malignant tumors. However, the biosynthetic pathway of camptothecin has remained elusive for five decades. Here, we fully map its entire biosynthetic route. We discovered five key missing enzymes (OpCAR, OpSDR11, OpCS, OpGH1, and OpSTR) via the combination of MALDI mass spectrometry imaging, single-cell RNA sequencing and co-expression analysis. Meanwhile, we demonstrated a free flavin mononucleotide triggered the non-enzymatic 6-5-6 to 6-6-5 fused-ring skeleton rearrangement, filling the last gap in camptothecin biosynthesis. Finally, we validated this identified pathway and achieved the de novo biosynthesis of camptothecin in Saccharomyces cerevisiae. These discoveries uncover the long-standing mystery underlying camptothecin and pave the way for manufacturing camptothecin and its derivatives through synthetic biology approaches.

synthetic biology↗

bioRxiv2026

OmicOS: A Comprehensive Omics Ecosystem Infrastructure and Agent System for the AI Era

Biology has accumulated a vast ecosystem of omics methods, but much of this ecosystem remains built for expert humans rather than scientific agents. Methods are scattered across Python packages, R/Bioconductor and CRAN workflows, command-line tools, incompatible data containers and implicit object states, making even routine analyses difficult for an AI system to choose, execute and verify reliably. Here we introduce OmicOS, a comprehensive omics ecosystem infrastructure and agent system that turns OmicVerse V2, an open-source omics community, into an executable foundation for agentic biology. OmicVerse V2 provides the community substrate: scalable AnnDataOOM-compatible rust backends, agent-friendly Python algorithms for single-cell, spatial, bulk and multi-omics analysis, interfaces to single-cell foundation models, and Python-native reconstructions of historically R-centred Bioconductor/CRAN-style workflows. OmicOS makes this substrate actionable by registering analytical functions as state-aware capability contracts, allowing agents to inspect live data objects, select valid methods, execute controlled workflows and record provenance. The result is not a fixed pipeline, but a programmable omics environment in which agents compose real analyses from verified community methods rather than inventing tools. Across external and purpose-built benchmarks, OmicOS ranked first among the evaluated systems, reaching 81.2% on BiomniBench. Adding OmicVerse to a minimal agent improved task completion by up to 34.2 percentage points with qwen-3.6-35b, and controlled ablations showed that the gains came from registry-grounded execution rather than from larger models, documentation retrieval or unrestricted tool exposure. The same infrastructure scaled to atlas-sized data, reproduced R-centred workflows in Python and converted external pathology software into agent-usable skills. In a discovery task starting from a whole-body spatial map and the term "Alzheimers disease", OmicOS composed a non-canonical workflow that integrated spatial expression, genetic association, eQTL and colocalization evidence to nominate a colon epithelial risk axis centred on PICALM, CD2AP and CR1. Together, OmicVerse and OmicOS define an open foundation for AI-era omics, showing how a community of biological methods can be transformed into a reliable, extensible and agent-operable system for discovery. HighlightO_LIOmicVerse 2.0 consolidates 694 methods spanning 11 omics domains into agent-callable high-level APIs. C_LIO_LIRebuildR automatically reconstructs and evolves R/Bioconductor methods as Python-native implementations under output-equivalence gates. C_LIO_LIOmicOS establishes a state-of-the-art omics agent harness, ranking first on general omics benchmarks across models and substantially improving the analytical capability of local open-source models. C_LIO_LICompositional use of ecosystem modules nominates a colon epithelial axis associated with Alzheimers disease risk. C_LIO_LIExternal algorithm packages supporting automatic iterative evolution can be integrated into the OmicOS ecosystem. C_LI

bioinformatics↗

bioRxiv2026

Reinforcement learning-driven unified generative framework for multi-objective RNA codon design

Current RNA codon design methods are limited by inefficient long-sequence processing and poor generalizability, often relying on a decoupled generate-or-optimize paradigm. We introduce RNARL, a reinforcement learning-driven framework that unifies sequence generation with multi-objective optimization. RNARL directly learns to generate high-performance sequences, effectively optimizing sequences over 3,900 nucleotides and demonstrating superior performance and universality across six species and five RNA types. RNARL thus establishes an effective and generalizable framework for RNA codon design. Finally, a user-friendly web platform is freely available to facilitate its application for RNA therapeutic design.

bioinformatics↗