Biology subjects

Yuan, Y.

Publications and source records attributed to Yuan, Y..

At least 19 recordsLinked to original sources

bioRxiv2026

Erythrocyte aging links hypocalcemia to impaired flow dynamics and oxygen transport through volume dysregulation

Red blood cells (RBCs) maintain their corpuscular volume through mechanotransducive ion-mediated processes in response to circulatory shear stress. However, with age, RBCs accumulate oxidative damage and undergo alterations in protein function, glucose metabolism and ion transport, raising fundamental questions about how volume homeostasis is preserved under mechanical load. Here, we investigated the differential responses of younger (YRBCs) and older RBCs (ORBCs) to fluid shear stress, with a focus on calcium-dependent regulation. We found that ORBCs exhibited a striking dependence on extracellular calcium to maintain volume stability. In calcium-depleted conditions, only ORBCs underwent pronounced post-shear swelling, with recovery only partially restored by ATP-dependent mechanisms. In contrast, calcium preserved volume homeostasis, identifying calcium-sensitive pathways as key compensatory regulators in metabolically compromised ORBCs. Notably, metabolically deficient YRBCs recapitulated the ORBC swelling phenotype, supporting a shift from ATP to calcium-dependent regulation with age. More importantly, shear-induced swelling impaired flow behaviour and oxygen delivery in ORBCs, linking defective volume control to microcirculatory dysfunction. Together, these findings potentially established calcium-dependent volume regulation as a critical adaptive mechanism in ageing RBCs and reveal a previously unrecognized hematological link to vascular dysfunction in hypocalcemia.

bioengineering↗

bioRxiv2026

Single-nucleus chromatin accessibility time course of excisional wound repair in rat skin

We present a time-series single-nucleus ATAC-seq dataset from full-thickness excisional wounds in Sprague-Dawley rats. Samples were taken from unwounded skin (normal control) and at 1, 2, 5, 10, and 19 days post-wounding, spanning the inflammatory, proliferative, and remodeling phases of skin repair. We extracted nuclei with a two-step tissue dissociation protocol and prepared libraries using the DNBelab C Series Single-Nucleus ATAC Kit, then sequenced on the MGI DNBSEQ-T1 platform. After stringent quality filtering, 37,555 nuclei passed quality control; the resulting chromatin accessibility profiles cover 23 cell types: fibroblast subpopulations, keratinocytes, vascular endothelial cells, immune cells, and skin appendage cell types. This genome-wide chromatin accessibility dataset is a resource for studying regulatory dynamics during excisional wound repair and for comparing results with other wound healing systems.

bioinformatics↗

bioRxiv2026

Longitudinal Metaproteomic and Lysine Crotonylation Remodeling Defines Clinical Response to Fecal Microbiota Transplantation in Ulcerative Colitis

Background Fecal microbiota transplantation (FMT) can induce steroid-free clinical remission in a subset of patients with ulcerative colitis (UC), but treatment responses remain heterogeneous and are not fully explained by taxonomic microbiome shifts. To define functional molecular features associated with FMT efficacy, we performed integrated longitudinal fecal metaproteomic and lysine crotonylation (Kcr) profiling. Methods We analyzed 87 fecal samples from 23 patients with active UC collected at baseline and at weeks 1, 8, and 12 after FMT. Patients were classified as responders (n = 13) or non-responders (n = 10) according to steroid-free clinical remission at week 12. Across samples, we identified an average of 3,682 proteins and 562 Kcr sites per sample. Results Responders showed greater metaproteomic and meta-crotonylomic complexity at week 12, sustained enrichment of Prevotella and Faecalibacterium, and eight directionally consistent baseline-altered proteins with candidate prognostic value. After FMT, responders exhibited coordinated microbial functional remodeling, including increased proteins involved in short-chain fatty acid (SCFA) metabolism, acyl-CoA-centered pathways, nutrient transport, and polysaccharide degradation, together with attenuation of host neutrophil-related inflammatory protein and Kcr signatures. Non-responders showed minimal or opposing molecular changes. Conclusions: These findings nominate longitudinal fecal metaproteomic and Kcr remodeling as functional molecular readouts of clinical response to FMT in UC and highlight lysine crotonylation as an underexplored metabolite-sensitive layer linking microbial metabolic restoration with host inflammatory resolution.

microbiology↗

bioRxiv2026

Application of Bayesian-Derived Gene Interaction Network and Subnetwork for Classification of Idiopathic Parkinson Disease with Biomarkers

Given the expenses associated with Parkinson Disease (PD), it is imperative to develop a method for accurate diagnosis before emergence of motor symptoms. While there are tests that can measure early onset of PD, these tests still rely on subjective input from patients and can thus be biased. Factors influencing their accuracy include environmental agents such as education and culture. Thus, blood-based transcriptomic signatures are a promising direction for biomarkers in PD that are minimally invasive and more objective than clinical evaluation. In this paper, we analyzed gene expression from the PPMI cohort. Using a Bayesian framework, a gene interaction network was constructed to enhance differential expressed genes (DEG) analysis through signal diffusion. In tandem with clinical traits, 13 genes were selected for optimal discrimination using a random forest model. Within the holdout test set, the model achieved an AUC of 0.972, outperforming contemporary models. The model performance demonstrates the potential of using blood-based transcriptomics to enhance clinical diagnosis of PD. Other genes recovered by network diffusion but missed by traditional DEG analysis are already validated as potential biomarkers in contemporary PD research. By combining transcriptomic signatures with clinical evaluation, accurate diagnosis of PD can happen before the major destruction of dopaminergic neurons, which is critical for effective intervention and treatment.

bioinformatics↗

bioRxiv2026

An AI-Enabled Translational Drug Discovery Framework for NALCN Channelopathy: From High-Throughput Screening to Therapeutic Candidate

Background: More than 90% of rare diseases lack an approved therapy, partly reflecting economic barriers in developing treatments for small patient populations. A common bottleneck is costly and resource-intensive hit-to-lead prioritization. Congenital contractures of the limbs and face, hypotonia, and developmental delay (CLIFAHDD) is an ultra-rare neurodevelopmental disease often caused by de novo gain-of-function variants in the sodium leak channel NALCN, with no approved treatments. Methods: We developed a cost and resource-efficient, AI-enabled translational drug discovery framework integrating phenotypic high-throughput screening (HTS), structural biology foundation models for hit-to-lead prioritization, and orthogonal experimental validation. We applied this framework for drug repurposing to identify NALCN inhibitors as therapeutic candidates for CLIFAHDD. We screened ~1,200 approved drugs using a fluorescence-based membrane potential assay in doxycycline-inducible HEK293 cells stably expressing the NALCN channelosome. Boltz-2 modeling predicted binding affinity and mode to prioritize hits. Lead repurposing candidates were selected by a multidisciplinary expert review panel. Lead candidates were evaluated using patch-clamp electrophysiology in HEK293 cells expressing wild-type or CLIFAHDD-associated NALCN variants, primary mouse hippocampal neurons, and a C. elegans NALCN gain-of-function model. Results: HTS identified 11 candidate NALCN inhibitors. Boltz-2 prioritized candidates based on predicted binding affinity and predicted binding within the NALCN pore, consistent with a direct pore-blocking mechanism. Aprepitant emerged as the lead candidate after multidisciplinary evaluation based on its predicted mechanism, pharmacologic properties, blood-brain barrier penetration, established safety profile, and regulatory approval in children and adults. Patch-clamp electrophysiology confirmed partial inhibition of NALCN-mediated sodium leak currents in HEK293 cells (IC50;=2.22 M), with comparable potency against Cav3.1 (IC50=1.63 M). In primary mouse hippocampal neurons, aprepitant reduced sodium leak currents, hyperpolarized resting membrane potential, and decreased spontaneous firing. Aprepitant rescued hyperlocomotor and dystonic phenotypes in a C. elegans gain-of-function NALCN model. There was notable concordance between AI-predicted hits, phenotypic screening hits, and experimental validation. Conclusion: These findings identify aprepitant as a lead therapeutic candidate for CLIFAHDD, and establish an AI-enabled translational drug discovery framework integrating phenotypic screening, structural modeling, and experimental validation for NALCN channelopathies. This scalable and resource-conscious strategy may accelerate hit-to-lead prioritization and drug discovery for rare and ultra-rare diseases lacking effective therapies.

neuroscience↗

bioRxiv2026

PPARγ brakes cytosolic DNA-induced macrophage inflammation by super-enhancer metabolic reprogramming

Cytoplasmic DNA sensing triggers intense innate immune responses that require tight regulation. PPAR{gamma} is a master metabolic regulator in macrophages, but its role in cytoplasmic DNA immunity remains unclear. Using poly(dA:dT) to mimic cytoplasmic DNA stimulation, we combined transcriptomics, epigenomics (PPAR{gamma}/H3K27ac ChIP-seq), and functional assays. Poly(dA:dT) induced a cytokine storm with MAPK activation and apoptosis/necrosis. PPAR{gamma} activation by rosiglitazone (RGZ) antagonized this response, reducing cell death and transcriptionally inhibiting MAPK genes (e.g., DUSP4/10). Mechanistically, PPAR{gamma} bound promoters/enhancers via PPRE motifs and distal intergenic regions, remodeling H3K27ac (reduced at inflammatory loci, increased at metabolic genes). RGZ induced de novo super-enhancers coupling PPAR metabolic programs (e.g., CPT2) with immune suppression by targeting TLR4, CEBPB, and PIM3. Thus, PPAR{gamma} acts as a brake in cytoplasmic DNA immunity, integrating metabolic reprogramming and inflammation suppression through a multi-level genomic/epigenetic program. This study highlights PPAR{gamma} as a key immune-metabolic integrator and therapeutic target for diseases driven by excessive cytoplasmic DNA sensing.

immunology↗

bioRxiv2026

Base-mismatched crRNA in CRISPR Enabled Universal One-pot Rapid Nucleic Acid Detection

CRISPR-Cas12a-based diagnostics are promising for field detection yet are often constrained by two step workflows or laborious optimization of one-pot reactions. Herein, we present a universal one-pot RPA/CRISPR-Cas12a strategy using base-mismatched crRNA in the seed region (positions 1-6). Introducing of single-nucleotide mismatches (complementary or error) into crRNA suppresses cis-cleavage activity, enabling amplification and detection in a single tube without PAM re-design or costly modified crRNA. Molecular dynamics (MD) simulations indicated that moderate potential energy (PE) variation correlates with improved screening performance, providing guidance for the selection of effective mismatches. This strategy successfully converted established two-step assays for Burkholderia pseudomallei and Brucella into one-pot formats, achieving detection of targets at 10 copies/L within 30 min with good specificity. Lyophilized pellets supported field testing on lithium-ion battery-powered devices, and experiments with simulated soil samples confirm the robustness of the approach. Together, this cost-effective, adaptable strategy can streamline RPA-CRISPR diagnostics for infectious disease surveillance.

molecular biology↗

bioRxiv2026

P-body sequestration of clock transcripts delays repressor synthesis to set circadian period in Drosophila

Negative-feedback oscillators require a delay between the accumulation of a repressor's mRNA and the action of its protein. In the circadian clock, this delay has been attributed largely to post-translational control of PERIOD (PER) stability and nuclear entry. The RNA-binding proteins shown to regulate per translation, ATAXIN2 and its partners, promote it, leaving open whether any step holds clock transcripts back before they are translated. Here, using time-resolved miniTurbo proximity labeling of endogenous PER across four phases of the circadian cycle in Drosophila clock neurons, we define a 252-protein PER proximitome that partitions into a nuclear arm and a cytoplasmic RNA-metabolism arm. A behavioral RNAi screen identified two P-body components, the DEAD-box helicase Me31B (DDX6) and the 5'-3' exonuclease Pacman (Pcm; XRN1), as strong regulators of circadian rhythms. Using single-molecule RNA-FISH, proximity RNA editing and ribosome profiling, we show that as per and tim transcripts accumulate, they localize to Me31B-labeled P-bodies and are poorly translated, most prominently at ZT12. Me31B knockdown disrupts P-bodies and releases per mRNA from them, causing PER to accumulate earlier and to ~2-fold higher levels, whereas Me31B overexpression delays PER accumulation and lengthens the free-running period by ~2 h. Knockdown of Pcm, in contrast, impairs clearance of per mRNA, sustaining PER and TIM accumulation, prolonging the repression phase and abolishing cycling of ~89% of rhythmic transcripts. Together, these findings identify P-body sequestration as a repressive step that delays repressor synthesis, and Pcm-dependent decay as required to end repression on time. Given the deep conservation of DDX6 and XRN1, RNP compartments may provide a conserved means of generating delay in circadian and other negative-feedback circuits.

cell biology↗

bioRxiv2026

Spatiotemporal atlas resolves the mesenchymal-epithelial stem cell axis governing human hair follicle aging and pathology

How adult stem cell populations coordinate tissue regeneration across physiological and pathological states remains unclear. Hair follicles contain epithelial and mesenchymal stem cells that codetermine hair regeneration. Here, we integrate a multimodal spatiotemporal atlas with functional validation to define the mesenchymal-epithelial stem cell dynamics governing human hair follicle homeostasis, aging, and androgenetic alopecia. Aging is characterized by passive depletion of hDSC-derived regenerative cues, whereas AGA involves a shift of hDSC toward an inhibitory signaling hub that promotes HFSC dysfunction. We further pinpoint hDSC-secreted PTN as a key regenerative factor that promotes hair growth and counteracts androgen-induced growth suppression. This study provides a holistic framework of adult stem cell crosstalk across physiological and pathological states, offering a mechanistic roadmap to reinstate tissue regeneration.

cell biology↗

bioRxiv2026

Democratizing Agentic Access to Bioinformatics and Biopharmaceutical Databases and Analyses with BioMCP-TS

Practitioners want to give AI agents direct access to biomedical databases and analyses, but today that means choosing between dependency-heavy skill packages, data-lookup-focused server tools, and closed vendor platforms. BioMCP-TS removes the choice: one open-source server that installs with a single pinned command and verifies its own setup. An agent working through BioMCP-TS can search and cross-reference 50+ bioinformatics, pharmaceutical, and patent databases (genes, variants, drugs, diseases, literature, clinical trials, patents, functional genomics, and structures) and, uniquely among open bioinformatics servers, run heavyweight analyses in-process as WebAssembly: Bioconductor differential expression and htslib genomics operations with no R installation, C toolchain, or containers on the host, plus read-only SQL over curated local databases. We demonstrate these strengths with seven practical cases spanning drug-target due diligence, translational intelligence, GWAS follow-up, dependency analysis, cohort genomics, and differential expression, ranging from a first federated lookup through target-disease landscapes to published-structure shortlists and reproducible RNA-seq results, each answered end-to-end within minutes. The server, recorded transcripts, and the full problem set are available at https://github.com/yeyuan98/biomcp-ts (npm: biomcp).

bioengineering↗

bioRxiv2026

Attenuated Salmonella-Mediated Delivery of GSDMD Potentiates PD-1 Blockade Therapy against Melanoma

Immunotherapy has emerged as a core therapeutic strategy for melanoma. Programmed death protein 1 (PD-1) is a critical immune checkpoint molecule that restrains host anti-tumor immunity, and therapeutic agents blocking the PD-1 signaling pathway have been widely deployed in clinical practice. Nevertheless, single-agent PD-1 blockade fails to elicit robust clinical responses in the majority of patients. Therefore, there is an urgent unmet need to develop combinatorial regimens capable of augmenting the anti-tumor efficacy of PD-1 inhibition. Gasdermin D (GSDMD), a pore-forming effector protein that orchestrates pyroptosis, exerts inherent anti-tumor activities upon overexpression. However, whether GSDMD can synergize with PD-1 blockade to enhance therapeutic outcomes against melanoma remains poorly defined. To address this question, we established an attenuated Salmonella engineered strain for targeted delivery of GSDMD, and further investigated the anti-melanoma therapeutic efficacy of combining this engineered bacterium with anti-PD-1 antibody via immunofluorescence staining, flow cytometry and other analytical approaches. Our in vivo results demonstrated that combinatorial treatment markedly suppressed melanoma progression in tumor-bearing mice relative to monotherapy with either GSDMD-expressing bacteria or anti-PD-1 antibody alone. Mechanistically, co-treatment upregulated intratumoral expression of GSDMD and the pro-apoptotic protein BAX, while simultaneously downregulating PD-1 expression. In addition, the GSDMD/anti-PD-1 combination significantly elevated the proportions of CD4 and CD8 T lymphocytes in both peripheral blood and splenic tissues, and facilitated robust tumor infiltration by these two T cell subsets. Compared with phosphate-buffered saline (PBS) and scramble control groups, combinatorial therapy promoted tumor infiltration of M1-type tumor-associated macrophages (TAMs) and repolarized TAMs away from the immunosuppressive M2 phenotype. Consistently, serum levels of the pro-inflammatory cytokines TNF- and IFN-{gamma} were markedly elevated following combined intervention. Collectively, this study verifies that attenuated Salmonella carrying GSDMD synergizes with anti-PD-1 antibody to elicit potent anti-tumor effects in melanoma-bearing mice by amplifying systemic and intratumoral anti-tumor immune responses, which provides a preclinical rationale for novel combinatorial therapeutic strategies against melanoma.

cancer biology↗

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

Juvenile AAV-Mediated MEF2C Gene Replacement Ameliorates Selected Phenotypes in Mef2c-Haploinsufficient Mice

MEF2C haploinsufficiency syndrome is a severe neurodevelopmental disorder for which no disease-directed treatment is available. We investigated whether neuron-directed adeno- associated virus (AAV) delivery of a functional MEF2C coding sequence during the juvenile period could modify disease-relevant phenotypes in mice heterozygous for a Mef2c exon 4 deletion. Transcript-level analysis identified a brain-enriched MEF2C isoform containing the 1 and {beta} regions (nMEF2C) and a skeletal-muscle-enriched isoform containing 2 but lacking {beta} (mMEF2C). Separate human-synapsin-driven AAV vectors encoding either isoform were administered at postnatal day 28. Control-treated Mef2c heterozygous mice retained baseline sociability but lacked social-novelty preference. Mice treated with either nMEF2C or mMEF2C displayed social-novelty preference and improved selected responses to a new social partner, whereas open-field effects were limited. nMEF2C replacement also corrected dark-phase wakefulness and non-rapid eye movement sleep abnormalities and modified selected state- dependent electroencephalographic ratios, without broadly changing absolute band amplitudes or social-contact electroencephalographic activity. Atlas-based whole-brain mapping revealed region-selective reductions in parvalbumin-immunoreactive profiles; direct statistical evidence of cellular rescue was confined to the secondary motor area after nMEF2C treatment. These findings show that selected MEF2C-dependent phenotypes remain modifiable during the juvenile period and support further optimization of MEF2C gene replacement with respect to isoform, dose, expression control, and cellular targeting.

neuroscience↗

bioRxiv2026

Benchmarking confidence estimation and rescoring for cyclic peptide-protein complex predictions

Cyclic peptides are a rapidly expanding class of therapeutics, but the reliability of deep-learning structure prediction for cyclic peptide-protein complexes has not been systematically evaluated. We assembled a curated benchmark of 111 nonredundant complexes spanning five cyclization chemistries and assessed two co-folding models, Boltz and Protenix, each generating 100 poses per target (22,200 total). Stratifying all poses by complex attributes, we found that disulfidecyclized peptides and small protein targets (200 or fewer target residues) were predicted significantly worse by both tools, with target size the largest and most consistent effect; overall accuracy nevertheless remained high (median top-pose DockQ of about 0.89, 96-98% of targets Acceptable or better), indicating that pose generation is rarely the bottleneck. Conversely, native model ranking scores correlated only moderately with pose quality (Spearman rank correlations of 0.53-0.66): approximately 12% of poses showed high model ranking score/confidence despite poor pose DockQ quality, and the highest-quality pose was not ranked first for nearly every target. We therefore augmented the native score with externally computed interface descriptors normalized by chain length, principally the per-residue density of inter-chain hydrogen bonds, in a gradient-boosted rescoring model evaluated under target-grouped cross-validation that prevents leakage, improving out-of-fold ROC-AUC for both tools, significantly so for Protenix. Together, these findings identify pose ranking, rather than pose generation, as the major limitation of current cyclic peptide-protein complex prediction and demonstrate that complementary structural features can improve confidence-based pose selection.

bioinformatics↗

bioRxiv2026

A Time-Resolved Single-Cell Atlas Reveals Infection-Status, Age-, and Sex-Dependent Immune Responses Drive Viral Disease Severity

The majority of mortality during viral infections occurs in older males; however, underlying mechanisms by which age and sex shape antiviral immunity and pathological inflammatory responses remain incompletely understood. Here, we performed time-resolved single-cell RNA sequencing across 16 conditions spanning age, sex, and four stages of influenza infection in mice, generating a high-resolution atlas. Aged mice demonstrate delayed antiviral and inflammatory responses in multiple myeloid cells, impairing viral clearance, which delays recovery. Similarly, endothelial cells from aging mice show prolonged inflammatory and antiviral gene signatures. Altered gene signatures in immune and endothelial cells result in a shift in endothelial-immune interactions in the aged lung. Further, the infection status of the cell is a major driver of transcriptional state, with infected myeloid cells exhibiting broad upregulation of genes, including interferon-stimulated, inflammatory, complement, and oxidative stress-related genes. To assess whether these age-associated transcriptional patterns are conserved in humans, we examined BAL cells obtained from healthy individuals and COVID-19 patients, and found that immune cells from aged COVID-19 patients had elevated antiviral and pro-inflammatory gene expression compared to cells from young patients. Our analyses of sex differences identified that multiple myeloid cell types in aged male mice, but not in young male mice, show persistent inflammatory responses at later stages of infection, a likely mechanism contributing to elevated mortality in older males. These data reveal how infection status of the cell, age, and sex interact to drive persistent inflammation and impaired resolution, providing a foundational resource for designing age- and sex-specific therapeutic strategies.

immunology↗

bioRxiv2026

Optimized Multiple Circular Sequence Alignment for Cyclic Peptide Motif Discovery

Head-to-tail (H2T) cyclized peptides are an increasingly important modality in drug discovery, combining high target affinity and selectivity with metabolic sta-bility. Because their underlying chemistry is still that of a linear amino-acid chain, their linear sequence representation is the native input format of main- stream sequence generative models now driving de novo peptide design (Slough et al., 2018; Rettie et al., 2025a;b). Discovering the conserved motifs responsible for a family's function across a library of such candidates requires a multiple sequence alignment (MSA). Because a cyclic peptide can be linearised at any residue, the alignment must additionally solve for the unknown rotation of each sequence, which is the multiple circular sequence alignment (MCSA) problem. However, leading MCSA heuristics (e.g. Ayad & Pissis, 2017) were tuned for the genomic regime (a few tens of long sequences) and become prohibitively slow on the cyclic peptide library regime (hundreds to thousands of shorter sequences). We close this gap by identifying quality-preserving optimisation opportunities, notably the library-scale preset tailored to short-sequence inputs (algorithmic details in Appendix A), and by adding an orthogonal multi-core and SIMD backend for further performance tuning, which gives near-linear thread scaling on the pairwise-comparison stage. We validate the pipeline on a library of 1,000 H2T cyclized peptides of length 18 targeting the oncoprotein Mouse double minute 2 human homolog (MDM2) produced by an internal peptide-design engine. In this practical setup, the optimised MCSA recovers the underlying positional motif of MDM2 binders at the same fidelity as the original MCSA implementation while running over 650x faster. Our optimised MCSA tool thus enables library-scale cyclic peptide sequence alignment and is publicly available at https://github.com/IVB-Generative-Biology/mars-turbo.

bioinformatics↗

bioRxiv2026

Large-scale functional coupling and computational modelling reveal frontotemporal hotspots in distributed network connectivity during working memory recognition, encoding, and retrieval

Working memory depends on coordinated activity across distributed brain networks, but how these interactions support recognition, encoding, and retrieval remains unclear. Here, we analyzed intracranial local field potentials from 1,652 sites in 36 epilepsy patients performing a naturalistic card-matching task. Phase-amplitude and phase-phase coupling revealed widespread condition-dependent interactions, with frontal and temporal regions emerging as recurrent network hubs. Event-related analyses showed reproducible within-trial coupling dynamics across participants. Directed phase transfer entropy further identified frequency-specific information flow, including frontal regions acting as a theta-band source and an alpha-band sink. Finally, a time-delay-embedded hidden Markov model identified task-locked latent states with condition-dependent occupancy trajectories and distinct coherence profiles. Together, these results suggest that working memory is implemented through dynamic, directionally organized large-scale networks, in which frontotemporal hubs coordinate flexible transitions among distributed neural states rather than operating in isolation.

neuroscience↗

bioRxiv2026

SPgen: Proteome-wide Spatial Proteomics generation using multi-modality foundation models

Spatial proteomics (SP) measures the spatial distribution of proteins within tissues, providing important insights into tissue function, disease, and therapeutic response. However, current SP technologies profile only a small fraction of the proteome and are limited by cost and measurement noise. Recent AI approaches enable predicting spatial protein expression from transcriptomic or histopathological data, but are typically restricted to paired datasets covering only tens of proteins, limiting their ability to generalize beyond experimentally measured protein panels. Here we present SPgen, a multi-modal foundation-model framework for proteome-wide spatial protein prediction. SPgen integrates protein sequences, functional annotations, transcriptomic profiles, and spatial information to learn transferable representations that enable inference beyond experimentally profiled proteins. Across diverse spatial proteomics datasets, SPgen accurately reconstructs measured spatial patterns, reduces measurement noise, and enables proteome-wide spatial prediction.

bioinformatics↗