Biology subjects

Li, Y.

Publications and source records attributed to Li, Y..

At least 109 records · Page 6Linked to original sources

bioRxiv2026

Visual speech enhances phoneme separability in human superior temporal gyrus

Visual speech, such as lipreading, facilitates spoken word recognition, but the neural mechanisms underlying audiovisual speech perception remain poorly understood. Visual cues may disambiguate fine-grained articulatory features during early perceptual stages or instead integrate with speech at more categorical, phoneme-level stages. To test how speech representations are modulated by visual input, we analyzed intracranial electroencephalography (iEEG) signals recorded from 12 epilepsy patients performing an audiovisual speech perception task. Participants perceived 16 monosyllabic words presented in auditory-only, visual-only, or congruent audiovisual formats. Words were constructed from four onset consonants (/b/, /g/, /m/, /n/) and four rimes (vowel nucleus and any coda consonants). We examined event-related potentials (ERP) in superior temporal gyrus (STG) and trained support vector machine (SVM) classifiers to decode word identity from neural activity at individual electrodes. Discrete and continuous confusion matrices captured complementary changes in classification accuracy and normalized inverse classification loss, a continuous proxy for classifier confidence. Decoding performance was hierarchically evaluated at the word, phoneme, and phonetic feature levels to determine the representations affected by visual speech. Congruent audiovisual speech increased classifier confidence for phoneme-level representations and improved decoding accuracy at both the word and phoneme level, without corresponding effects on phonetic features. Time-resolved analyses further revealed earlier successful decoding for audiovisual than auditory-only speech, with audiovisual enhancement primarily observed for onset consonants rather than rimes. Together, these findings suggest that visual speech sharpens primarily categorical phoneme representations in STG, with accelerated speech processing and improved word recognition emerging as downstream consequences of phoneme-level enhancement.

neuroscience↗

bioRxiv2026

A SUN1-ApoD feedback loop promotes cellular aging via microtubule-nuclear mechanotransduction

Alterations in nuclear envelope proteins such as SUN1 and prelamin A are known hallmarks of cellular aging, but how they cause aging phenotypes beyond the expressing cells is unknown. With a tissue culture version of parabiosis, we found that aged fibroblasts secrete an activity that induces aging-related polarity defects in young fibroblasts. Secretomics identified the factor as apolipoprotein D (ApoD), a circulating protein whose levels are known to rise with age. ApoD increased SUN1 levels and disrupted polarity in fibroblasts via SUN1-promoted coupling of microtubules to the nucleus. In turn, elevated SUN1 enhanced ApoD expression and secretion, forming a feedback loop that promotes acquisition of aged phenotypes. In mice, ApoD induced aging-related phenotypes in muscle. Elevated ApoD expression required coupling between SUN1, its outer nuclear membrane binding partner nesprin-2, and microtubules. Broader analysis revealed that this pathway regulates hundreds of genes. These findings define microtubule-nuclear coupling as a direct mechanotransduction pathway controlling gene expression and promoting aging-associated phenotypes.

cell biology↗

bioRxiv2026

Proteomic and Kinetic Analyses Reveal Discordant Apolipoprotein Turnover and Support a Revised Model of Human Lipoprotein(a) Metabolism

Objective: Lipoprotein(a) [Lp(a)] is a causal risk factor for atherosclerotic cardiovascular disease composed of apolipoprotein(a) [APO(a)] covalently linked to apolipoprotein B100 (APOB). Although plasma Lp(a) concentrations are largely genetically determined, the mechanisms governing Lp(a) metabolism after particle assembly remain poorly understood. We sought to define the Lp(a) proteome and determine the metabolic behavior of APO(a) and APOB within circulating Lp(a) particles using integrated proteomic, kinetic, and imaging approaches. Approach and Results: Sixteen healthy adults underwent stable isotope tracer studies with 2H3-L-leucine and 2H5-glycerol. Lp(a) particles were isolated by APO(a)-specific immunoprecipitation for high-resolution liquid chromatography-mass spectrometry and kinetic analyses, and extracellular vesicles (EVs) were characterized by imaging flow cytometry and super-resolution microscopy. Proteomic analysis identified 92 proteins associated with immuno-isolated Lp(a), enriched in pathways related to immunity, coagulation, and atherogenesis. Lp(a)-APO(a) exhibited a mean fractional clearance rate of 0.04 pools/day, whereas Lp(a)-APOB cleared approximately sevenfold faster (0.25 pools/day), independent of plasma Lp(a) concentration or APO(a) isoform size. Both APO(a) and APOB were detected on circulating EVs, suggesting that EVs may contribute to post-secretory Lp(a) particle remodeling. Conclusion: These integrated human studies demonstrate marked discordance between APO(a) and APOB turnover within circulating Lp(a) particles, challenging the prevailing assumption that both apolipoproteins behave as a single metabolic unit after Lp(a) assembly and supporting a revised model of human Lp(a) metabolism.

physiology↗

bioRxiv2026

FFPERescuer: deep unsupervised domain adaptation for the reconstruction of gene expression profiles derived from formalin-fixed paraffin-embedded samples

Formalin-fixed paraffin-embedded (FFPE) tumor tissues often suffer from RNA degradation, posing a long-standing challenge for reliable transcriptomic profiling. Here, we propose FFPERescuer, a deep learning framework employing unsupervised domain adaptation, to rectify distorted gene expression data. FFPERescuer comprises a partial encoder that maps a small subset of genes to high-level representations and a decoder to reconstruct full gene expression profiles. On simulated data with varying noise levels, FFPERescuer faithfully recovered gene expression profiles, achieving high Pearson correlation coefficients (PCCs > 0.85) with the ground truth. In FF-FFPE-matched cohorts, FFPERescuer significantly enhanced expression profile concordance, with average PCCs increased by 23% (P < 0.05). Applying to cancer subtyping, FFPERescuer improved classification accuracy from 67% to 92%, recapitulated subtype-specific biological properties lost in the FFPE-derived data, and enhanced survival associations. Our studies provide a powerful framework for reliable transcriptomic profiling from FFPE-archived tumor samples that are widely available in the clinic.

bioinformatics↗

bioRxiv2026

Activity-dependent myelination in prefrontal circuits signals the offset of infantile amnesia

Human and non-human infants can form memories for events, but these memories are not successfully consolidated into remote memory. While the neurobiological basis of this phenomenon-known as infantile amnesia-remains unclear, it is hypothesized that the neural circuits required for successful consolidation are insufficiently mature. Here we find that heightened activity in the prelimbic cortex of developing mice triggers a sequalae of maturational steps that culminates in adult-like memory persistence: Activity-dependent increases in brain-derived neurotrophic factor (BDNF) promote myelination of prelimbic circuits via activation of tyrosine kinase receptor B (TrkB) receptors on oligodendrocyte precursor cells (OPCs). Inhibiting any of these steps within a critical developmental window delays the offset of infantile amnesia, whereas promoting this sequalae results in the precocial emergence of memory persistence. Similar to critical periods in sensory cortices, our results indicate that developmental myelination is required for proper circuit maturation and emergence of adult-like memory function.

neuroscience↗

bioRxiv2026

TLR-mediated activation of synovial fibroblasts from osteoarthritis patients promotes chondrocyte dysfunction

Objective: Toll like receptor (TLR) activation by cartilage derived damage-associated molecular patterns contributes to osteoarthritis (OA) pathogenesis, but the role of synovial fibroblasts in this process remains incompletely understood. We investigated the TLR responsiveness of human OA synovial fibroblasts and determined how TLR activation influences fibroblast-mediated regulation of chondrocyte function. Design: Primary synovial fibroblasts isolated from OA patients were characterized for TLR expression and stimulated with agonists targeting TLR1/2 to TLR9. Inflammatory mediator production, matrix metalloproteinase expression, mitochondrial respiration, and transcriptomic responses were assessed. Functional effects on cartilage homeostasis were evaluated using autologous chondrocyte spheroid co-cultures. Results: Approximately 50% of cells within OA synovial membranes were fibroblasts (CD45-CD31-PDPN+). Synovial fibroblasts expressed multiple TLRs at both the transcript and protein levels. Activation of TLR1/2, TLR4, TLR5, and TLR2/6 induced robust expression of IL-6, IL-8, G-CSF, MMP3, and MMP10, with minimal effects on mitochondrial respiratory function. Transcriptomic analysis revealed activation of inflammatory pathways together with enrichment of antigen presentation and protein translation programs. In autologous co-culture, TLR1/2-activated synovial fibroblasts promoted inflammatory and catabolic gene expression, suppressed anabolic gene expression, and impaired chondrocyte spheroid growth. Conclusions: Human OA synovial fibroblasts are highly responsive to TLR activation and acquire a pro-inflammatory, cartilage-degrading phenotype that directly impairs chondrocyte homeostasis. These findings identify synovial fibroblasts as key effectors of innate immune signalling within the OA joint and support targeting shared TLR-mediated pathways in joint-resident cells as a potential disease-modifying therapeutic strategy.

immunology↗

bioRxiv2026

Pyroptotic cell-derived factors promote osteoclast differentiation

The NLRP3 D301N substitution, the murine ortholog of the human D303N variant, causes NLRP3 constitutive activation (NLRP3CA) and inflammasome assembly. Here, we show that priming signals induced by lipopolysaccharide (LPS) are sufficient to trigger GSDMD-dependent pyroptosis in NLRP3CA-expressing bone marrow-derived macrophages (BMDMs), but not in wild-type (NLRP3WT) cells. NLRP3CA mice exhibit elevated IL-1{beta} secretion and LDH release in bone marrow supernatants under both basal and LPS-challenged conditions. Other cytokines (e.g., TNF-) increased comparably between genotypes, indicating that NLRP3CA specifically amplifies inflammasome-dependent responses. Conditioned medium (CM) from LPS-treated NLRP3CA BMDMs significantly enhanced osteoclast (OC) differentiation in vivo and in vitro compared to NLRP3WT CM, as did bone marrow supernatants from NLRP3CA mice. This osteoclastogenic activity was largely independent of IL-1{beta}, as demonstrated by experiments using IL-1{beta}-deficient NLRP3CA BMDMs, IL-1{beta}-neutralizing antibodies, and genetic deletion of the IL-1 receptor. The osteoclastogenic factors were primarily soluble proteins, as heat inactivation and proteinase K digestion markedly reduced OC formation. Extracellular vesicles (EVs) from NLRP3CA CM modestly promoted OC differentiation, but EV-depleted supernatants retained full activity, indicating that the primary mediators are soluble proteins not carried within EVs. Together, these findings establish NLRP3CA as a well-controlled model for studying pyroptosis and demonstrate that pyroptotic cells release soluble proteins that drive osteoclastogenesis independently of IL-1{beta}.

cell biology↗

bioRxiv2026

Conservation of Nuclear Receptor DAF-12 with Evidence for Ligand Diversification in Plant-Parasitic Nematodes

Plant-parasitic nematodes cause extensive crop losses that threaten global food security and create substantial economic hardship, yet strategies for their control remain limited. Although plant-parasitic nematodes have life cycles that mirror those of their free-living and animal-parasitic relatives, comparatively little is known about the signals that drive life-cycle progression. Dafachronic acids (DAs) are steroid hormones that control life cycle by activating DAF-12 nuclear receptors in Caenorhabditis elegans and animal-parasitic nematodes. To define whether an analogous pathway exists in plant-parasitic nematodes, we investigated the responsiveness of DAF-12 orthologs from six plant-parasitic nematode species to {Delta}7-DA, an endogenous DAF-12 ligand in C. elegans and several animal-parasitic nematodes. We find that whereas some plant-parasitic nematode receptors are activated by {Delta}7-DA, others are not. Structural modeling of all six ligand-binding domains, together with docking of {Delta}7-DA into each model, indicates that responsiveness is governed by a combinatorial ligand-recognition network surrounding the ligand carboxylate rather than by any single contact residue. We further demonstrate that egg and second-stage juvenile extracts from the soybean cyst nematode, Heterodera glycines, lack detectable DA but contain distinct lipophilic fractions that activate its DAF-12 ortholog. Together, these findings suggest that many plant-parasitic nematodes utilize endogenous DAF-12 ligands that are structurally distinct from DA. Uncovering modulators of this pathway could lead to new strategies to combat nematode-driven crop loss.

molecular biology↗

bioRxiv2026

A Multi-Omics Study of Cognitive Resilience to Alzheimer's Disease

Alzheimers disease (AD) is the most severe form of dementia. While significant efforts have been made to identify AD risk factors and develop therapeutics, much less attention has been given to neurobiological factors and molecular signatures associated with cognitive resilience. Substantial evidence suggests that modifiable factors, such as physical and mental activity, may contribute to cognitive reserve and resilience. To this end, we developed a resilient cohort of paired genetics, transcriptomics and proteomics data from the prefrontal cortex of 282 cognitively normal older adults. We conducted extensive preprocessing and rigorous quality control (QC), including QC on sequence metrics and sample matching across multiomics. This cohort can be leveraged for identification of molecular signatures of cognitive resilience to AD and is freely shared with the research community.

genomics↗

bioRxiv2026

OmniSyn unifies target-aware molecular generation and optimization within a synthesis-native LLM framework across the human proteome

Designing target-specific bioactive molecules with actionable synthesis routes for the human proteome holds enormous potential for expanding therapeutic discovery, but remains a challenge. Existing target-aware generative models often depend on protein structures and generate molecules before assessing synthetic feasibility. Here we present OmniSyn, a protein-sequence-conditioned Mixture-of-Experts (MoE) language model that couples a task-conditioned interaction module with a synthesis-action decoder to generate molecules with explicit synthesis traces, thereby unifying de novo ligand generation, synthesizability projection and hit-to-lead (H2L) optimization within synthesis-traceable chemical space. OmniSyn is pre-trained with self-distillation and post-trained with task-specific reinforcement learning (RL) to adapt expert routing and optimize molecular properties across design modes. On unseen protein targets from MolGenBench, a real-world drug-discovery benchmark, OmniSyn achieves state-of-the-art performance across de novo design and H2L optimization, including target-awareness and hit-rediscovery metrics, despite relying only on protein sequences rather than three-dimensional (3D) pocket structures. By embedding synthesis planning into the design process, OmniSyn shifts molecular generation from a generate-then-filter paradigm toward design-with-synthesis paradigm, transforming virtual predictions into experimentally actionable candidates. Independent AiZynthFinder evaluation yielded retrosynthetic success rates of 68.47% for de novo generation and 71.92% for H2L optimization, improving over the strongest baselines by 61.3% and 184.2%, respectively, and supporting the synthetic feasibility of OmniSyn-generated molecules. In synthesizability projection, OmniSyn further converts outputs from external generative models into close analogues with improved retrosynthetic feasibility while preserving molecular similarity. Having established strong benchmark performance and external retrosynthetic feasibility, we next applied OmniSyn at human-proteome scale, spanning more than 21,000 targets. Rapid sequence-conditioned sampling enabled the construction of, to our knowledge, the largest human-proteome-scale generative virtual library, comprising 2.7 billion target-specific molecules, each accompanied by model-derived synthesis traces and target-specific prioritization scores. By enabling scalable target-specific molecular design with synthesis-aware generation across the human proteome, OmniSyn opens new opportunities for exploring previously inaccessible therapeutic targets, including those lacking experimentally resolved structures.

bioinformatics↗

bioRxiv2026

A pan-cohort transcriptional landscape of breast cancer maps subtype and microenvironmental programs

Breast cancer comprises heterogeneous transcriptional states that are incompletely captured by discrete clinical or molecular subtype labels. To visualize this heterogeneity in a unified framework, we integrated bulk RNA-seq data from 2,284 patient samples across 13 studies using 18,089 protein coding genes, a harmonized processing pipeline, batch correction, consensus clustering and PaCMAP dimensionality reduction to construct an interactive breast cancer transcriptional landscape. Consensus clustering identified five major regions, which were annotated using PAM50 scores calculated for each sample: Luminal A, Luminal B, HER2 enriched, and two basal associated clusters. The basal clusters separated into an immune rich region marked by T cell-inflamed, tumor-associated macrophages (TAM), and low-purity signatures, and a cell-cycle driven region enriched for proliferation and DNA replication programs. Overlay of marker genes, pathways, kinases, neuronal like signaling programs, cancer associated fibroblasts (CAF) states, and TAM programs revealed spatially organized subtype biology and microenvironmental heterogeneity. Finally, projection of therapy associated resistance signatures identified landscape regions linked to predicted resistance to HER2-targeted therapy and hormone receptor directed endocrine therapies. By enabling interactive exploration of transcriptional states, marker genes, pathways, and therapeutic response programs, this resource provides a community framework for biomarker discovery in breast cancer.

bioinformatics↗

bioRxiv2026

Frontal stimulation reshapes SCAN connectivity and improves motor function in Parkinson's disease

Parkinson's disease is a neurodegenerative disorder with characteristic motor deficits of bradykinesia and impaired coordination, and non-motor deficits affecting cognitive abilities such as attention and executive functions. Recent evidence suggests that some canonical motor deficits in Parkinson's disease (e.g., freezing of gait) are associated with abnormalities in frontal cortical processing typically thought to underlie executive functions such as cognitive control. The evidence from human research linking frontal cognitive disruption and motor deficits in people with Parkinson's disease is correlational. Individuals with cognitive impairments often also display gait and balance deficits, but causal evidence remains lacking. We combined non-invasive brain stimulation with functional MRI (fMRI) to ascertain whether dorsolateral prefrontal cortex (DLPFC) exerts a causal influence modulating motor function and to test the hypothesis that an excitatory stimulation protocol applied to a cognitive control network can improve motor performance in Parkinson disease. Participants with Parkinson's disease performed a precision force-tracking task shown to track with gait impairment while undergoing functional neuroimaging scans. On half of trials, participants were required to simultaneously perform a cognitively demanding 2-back working memory task to tax attentional systems. Following the initial baseline scan, participants returned for three separate counterbalanced sessions that combined transcranial magnetic stimulation (TMS) in the form of theta burst stimulation (TBS) with fMRI. Just prior to task performance, participants received either excitatory (intermittent TBS) or inhibitory (continuous TBS) stimulation protocols over right DLPFC. In the third session, participants received stimulation to a control site outside the cognitive and motor areas implicated in our task. Our within-subjects experimental design allowed us to examine the effects of stimulation on motor behavior, the influence of cognitive load on motor behavior, and the neural systems supporting cognitive control and motor execution. Consistent with our hypothesis, excitatory stimulation of DLPFC improved motor performance in a precision force-tracking task as indexed by reduced tracking error and smoother movement execution. These effects emerged preferentially under dual-task conditions when cognitive demands were high. Our results show that this improved motor performance coincides with reductions in functional connectivity not only between the stimulation site and primary motor cortex, but also within the recently described somato-cognitive action network (SCAN) thought to support action planning and whole-body movement coordination. We conclude that frontal cognitive control systems play a causal role in modulating movement in Parkinson's disease and suggest that targeting these systems with non-invasive interventions may enhance motor function.

neuroscience↗

bioRxiv2026

A Hymenoptera-restricted gene mediating ant castes co-opts deeply conserved machinery to control organ size

Lineage-specific genes are widespread and have been implicated as phenotypic innovation inducers, but how they acquire complex developmental functions remains poorly understood. Ant queens and workers develop dramatically different organ sizes from identical genomes under juvenile hormone (JH) control, yet the molecular effectors translating JH signalling into caste-specific organ growth remain unknown. Here we identify torch, a Hymenoptera-restricted gene, as the most consistently gyne-biased and JH-responsive gene across 68 ant species. Knockdown of torch in virgin queens of Monomorium pharaonis produces a worker-like, multi-organ growth-restricted phenotype. Mechanistically, torch harbours an E-box-like motif activated by the JH receptor Gce-Tai and acts as a GA-repeat-binding transcription factor that regulates Hippo signalling, the deeply conserved organ-size control pathway in animals. Expressing torch heterologously in mice and a growth-restricted Drosophila background shows that the gene retained its general growth-promoting activity across more than 700 million years of animal evolution in lineages that lack the gene, establishing that its function is mediated through conserved rather than ant-specific machinery. A lineage-specific gene can therefore acquire complex morphogenetic function by co-opting ancient organ-size circuitry, providing a general route by which novel genes can drive phenotypic innovation.

evolutionary biology↗

bioRxiv2026

A patient-centric therapeutic paradigm uncouples prostate cancer suppression from systemic metabolic collapse

The clinical benefits of cancer therapies are often compromised by the tolerable adverse effects that impair systemic organismal health and may evolve into latent life threats. Here, we identified profound abiraterone-induced but androgen-independent metabolic perturbations in prostate cancer patients and developed Lifehug-9892 to balance tumor therapy with systemic metabolic homeostasis. By integrating population cohorts with high-resolution metabolomics, we demonstrate that abiraterone induces profound systemic lipidomic dysregulation, characterized by the massive, pathological accumulation of desmosterol. Abiraterone inhibits but stabilizes DHCR24, leading to a metabolic trap in patients showing elevated levels of both desmosterol and cholesterol. Desmosterol accumulation is highly lipotoxic, potently triggering endothelial cell senescence and necrosis, macrophage foam cell formation, murine atherosclerosis, and hepatic senescence. To mechanistically uncouple and therapeutically rescue this systemic metabolic collapse, Lifehug-9892 was rationally designed to selectively retain on-target CYP17A1 inhibition while completely sparing DHCR24 function. Lifehug-9892 maintains potent tumor-suppressive activity while fully preserving the desmosterol-cholesterol metabolic axis and preventing systemic cardiovascular and hepatic damage. Our study uncovers a critical mechanistic link between drug-induced metabolic dysregulation and organismal health in cancer patients, providing a biochemical framework for developing patient-centric targeted therapies that preserve host homeostasis.

cancer biology↗

bioRxiv2026

The circadian system is affected by Alzheimers disease independently from amyloid beta deposits

Circadian disruption, notably sleep disturbances, serves as an early indicator of Alzheimers disease (AD), preceding cognitive symptoms like memory loss. The suprachiasmatic nucleus (SCN) governs biological rhythms and receives direct retinal input via melanopsin-expressing retinal ganglion cells (mRGCs) to synchronize with environmental light cycles. The anatomical and functional basis for circadian disruption in AD remains unclear. Here, we explored the multi-level relationships between gene expression, the SCN connectome, and regulations of sleep and circadian rhythms in the APP/PS1 mouse model. The sleep architecture of APP/PS1 mice displayed significantly reduced rapid eye movement sleep (REM), associated with a reduced daily core body temperature amplitude and locomotor hyperactivity. Lastly, APP/PS1 mice showed an impaired response to acute light pulse stimulation and present hyperactivity of mRGCs at a young age and hypoactivity of these cells at older ages. These physiological functions are known to be, at least in part, regulated by the SCN, the main target of mRGCs. We noted several modifications in SCN connectomics using serial blockface electron microscopy (SBEM), including a reduction of the dendro-dendritic chemical synapse (DDCS) network that receives a large part of the retinal input and is thought to be crucial for synchronicity between SCN neurons. In addition, we observed multiple signs of dystrophy, including modifications of the shape of dendrites and cell soma, accumulation of aggregated lysosomes, and swelling of axons. At the same time, we investigated the changes in gene expression using spatial transcriptomics. The SCN presents changes in the expression of genes associated with synapse formation, cell adhesion, and neurite growth. These results suggest that, despite the absence of amyloid plaques in the ventral hypothalamus, the SCN of APP/PS1 mice still undergo profound gene expression changes, impacting connectomics and physiological functions. Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=157 SRC="FIGDIR/small/744599v1_ufig1.gif" ALT="Figure 1"> View larger version (39K): org.highwire.dtl.DTLVardef@ceedb0org.highwire.dtl.DTLVardef@156cfaaorg.highwire.dtl.DTLVardef@5bc262org.highwire.dtl.DTLVardef@36df4d_HPS_FORMAT_FIGEXP M_FIG C_FIG

neuroscience↗

bioRxiv2026

A transcriptomic and spatial map of serotonin autoreceptor expression in Drosophila

Serotonin is an evolutionarily ancient neurotransmitter that modulates an array of behaviors such as mood, sleep, and appetite across species. Serotonin acts primarily by binding to serotonin receptors, which are expressed in post-synaptic neurons (heteroreceptors) and serotonergic neurons themselves (autoreceptors). Serotonin autoreceptors modulate serotonergic tone, the foundational principles of which have been excellently demonstrated in vertebrate and invertebrate models. However, many aspects of the mechanisms and contexts in which this modulation occurs are still unclear. Drosophila melanogaster is a powerful model organism that can provide unique insights into autoreceptor function by the ability to perform precise spatial and temporal genetic manipulation with structural and functional readouts / behaviors of serotonin systems. However, a systematic characterization of serotonin autoreceptor expression in Drosophila has not been conducted. Here we use single-cell sequencing and genetic labeling to show that all five serotonin receptors are expressed in serotonergic neurons and map their expression at both the larval and adult stages of development. This is the first evidence of 5-HT2A and 5-HT7 expression in serotonergic neurons in any organism. Moreover, the unique combinations of autoreceptor expression in specific neuronal clusters will aid in the development of novel hypotheses for autoreceptor function, and demonstrates the utility of Drosophila as a model organism to study the function of serotonin autoreceptors.

neuroscience↗

bioRxiv2026

TigerAI: An AI-powered genetic evidence platform to support clinical development

Genetic evidence is a major determinant of clinical success in drug development, yet its aggregation has long relied on laborious human curation. Large language models (LLMs) have the potential to rapidly synthesize knowledge across biomedical resources, providing a route to scalable AI-driven genetic evidence generation. Here we develop a novel domain-grounded instruction framework to systematically evaluate GPT-5 for producing genetic evidence relevant to clinical trial success. Using 13,022 target-indication pairs from a comprehensive drug development database, we benchmark LLM-derived evidence against a recent exhaustive human expert-curated study. We find that GPT-5 yields genetic evidence that is at least as informative as expert curation for inferring clinical success, while substantially expanding coverage relative to traditional curation resources. Building on these results, we introduce TigerAI (https://tigerai.bio/), a dual-purpose platform for AI-powered genetic evidence that (i) benchmarks emerging state-of-the-art LLMs and (ii) provides an accessible service for querying reliable AI-generated genetic evidence. These contributions outline a practical, domain-grounded pathway for integrating AI-powered genetic evidence into drug development pipelines and for realizing the potential of LLMs to inform clinical success.

genetics↗

bioRxiv2026

Shared and idiosyncratic coding regimes coexist in macaque IT

Population-level measurements portray object representations in primate inferotemporal cortex (IT) as smooth, low-dimensional, and predictable by deep neural networks (DNNs). However, it remains unclear whether this structured population-level picture is representative of the full diversity of its constituent neurons. Here, we used Neuropixels 2.0 probes to record large populations of well-isolated units from an fMRI-localized face patch in macaque anterior IT while monkeys viewed more than 3,000 natural images. Single-unit responses were substantially sparser and more heterogeneous than multi-unit activity (MUA). Sparse neurons collectively formed a higher-dimensional code, supported efficient image identification, and responded later than broadly tuned neurons. Partly independently of sparseness, some neurons exhibited reliable feature randomness: their stimulus preferences were reproducible across repeated presentations but discontinuous across DNN feature spaces, rather than reflecting trial-to-trial response variability or noise. These neurons were virtually uncorrelated with the surrounding population despite being located within the same face patch. By contrast, MUA responses were denser, more correlated, lower-dimensional, and better predicted by DNNs. Together, these findings suggest a dual coding architecture in anterior IT: shared low-dimensional structure supports category generalization, whereas sparse and reliably feature-random single-neuron responses expand the representational space and support efficient identification of individual visual inputs. Response sparseness and reliable feature randomness may therefore constitute fundamental computational resources for object recognition.

neuroscience↗