Biology subjects

Singh, A.

Publications and source records attributed to Singh, A..

At least 55 records · Page 3Linked to original sources

bioRxiv2026

Targeting Autophagy Accelerates Intestinal Repair after Acute Ionizing Radiation

Radiation exposure induces systemic and cellular damage, contributing to acute radiation syndrome and long-term effects such as premature aging and carcinogenesis. At the cellular level, radiation triggers apoptosis, mutation, and transformation through oxidative damage and activation of pathways including ER stress-mediated autophagy. Autophagy plays a context-dependent dual role in stressed cells, but its contribution to intestinal recovery after acute radiation remains unclear. Here, we evaluated combinatorial radiomodification using gamma radiation (8 Gy) and autophagy modulators in whole-body irradiated C57BL/6 mice (8-10 weeks old, n = 10). Mice were treated with autophagy inducers or inhibitors and euthanized at 3-, 8-, and 30-day post-irradiation. The jejunal-ileal region was analyzed via antioxidant assays, immunoblotting, H&E staining, and immunohistochemistry. Radiation significantly altered oxidative stress and autophagy markers, including increased LC3-II and decreased SQSTM1/p62. Autophagy induction enhanced intestinal proliferation (as measured by Ki-67), whereas inhibition impaired regeneration. Rapamycin pretreatment improved survival and reduced markers of intestinal injury following 8 Gy total body irradiation (TBI), whereas chloroquine exacerbated several injury-associated parameters. Overall, our findings suggest that targeted modulation of autophagy is a promising strategy for alleviating radiation-induced gastrointestinal injury and provide mechanistic insights relevant to therapeutic development.

molecular biology↗

bioRxiv2026

Integrated multi-omics profiling of gut bacterial extracellular vesicles links cargo composition to host transcriptional responses

Bacterial extracellular vesicles (bEVs) enable gut microbiota to deliver bioactive cargo to host cells, yet the specific bacterial producers have not been systematically identified. Here, we profiled stool-derived bEVs from healthy individuals using metaproteomic profiling, revealing that vesiculation is widespread across gut bacterial phyla. We selected a subset of vesiculating species, showing that bEVs localize to distinct tissues in vivo, including extraintestinal sites such as lung, liver, kidney, and bone, suggesting roles beyond the gastrointestinal tract. We profiled intestinal epithelial cells and macrophages after endocytosing bEVs from various species, uncovering pronounced, cell-type-specific responses. For example, Bacteroides fragilis bEVs promote anti-inflammatory mitochondrial-telomeric regulation, while a set of commensal-derived bEVs contribute to epithelial survival and structural renewal. By combining proteomic, lipidomic, and metabolomic cargo profiling with transcriptional output, we find that specific Bacteroidota-derived protein cargo activates cytoprotective stress defenses while attenuating inflammatory signaling. Together, these findings establish a multi-layered comparative atlas of bEV composition, uptake, and host response, providing a framework for understanding bEV-mediated microbiome-host communication.

microbiology↗

bioRxiv2026

AI-enabled Spatial Profiling of Circulating Tumor-Immune Ecosystems Predicts Patient Outcomes Across Cancers

Circulating tumor cells (CTCs) and immune cells form dynamic multicellular ecosystems in blood, but their spatial organization and clinical relevance have not been systematically characterized. We developed the Cell and Cluster Identification Program (CCIP), an artificial intelligence-based framework that analyzes routine multiplex immunofluorescence blood scans to segment cells, identify CTCs and five immune lineages with high accuracy, and quantify multicellular clusters and tumor-immune interactions. Applying CCIP to 2,693 blood scans from 1,399 patients, we profiled over 60 million cells (>7 million multi-cell clusters) and linked imaging-derived features to patient outcomes. Correlated with circulating-tumor DNA mutation burdens, a 14-feature image model predicted overall survival in breast cancer, outperformed clinicopathologic variables and CTC enumeration, and generalized to prostate cancer. Prognostic imaging signatures were also associated with therapy response-related progression-free survival as well as with single-cell RNA sequencing-derived immune suppression states, connecting circulating tumor-immune architecture with systemic immune dysfunction.

cancer biology↗

bioRxiv2026

Visualizing immunoreceptor forces and their effects in vivo

Immunoreceptors experience forces that modulate their activities; however, demonstrating this in vivo has been limited by technical challenges. As a first step toward meeting this challenge, we adapted a synthetic Notch (SynNotch) receptor system to report force transmission through immunoreceptors in vivo by replacing the native ligand-binding domain with a receptor-specific antibody and rewiring Notch signaling to drive EGFP or luciferase expression. We expressed SynNotch on Jurkat cells targeting CD40 or T cell receptor (TCR) and characterized their activation in coculture with B or T cells, defining the requirements, optimal conditions, and kinetics of activation. Using complementary mechanobiology approaches, we quantified the exogenous force required for reporter activation and verified that activation depends on forces generated by receptor-expressing sender cells rather than SynNotch-expressing receiver cells. By implanting sensors and targets into immunocompromised mice, we visualized mechanically activated reporter expression on CD40 and TCR-targeting SynNotch cells in vivo. Furthermore, CD40 and TCR signaling was amplified when the receptor bore force against mechanical support from immobilized ligand, indicating that force functions as biologically relevant co-stimulus. Together, our results establish mechanically activated SynNotch reporters as a useful strategy for detecting receptor-associated mechanical signaling across 2D coculture, 3D organoid, and in vivo systems.

bioengineering↗

bioRxiv2026

Co-existence of Modularity and Anti-modularity in the Functional brain connectomes

The brain requires coordination among different regions to execute cognitive tasks, which may involve both positive- and negative-correlations. The topology of these correlations may indicate the mechanism underlying brain functioning in a given state. Here, we study changes in the functional connectomes (FCs) of both the positive and negative-correlations across various cognitive task states relative to the resting state, using publicly available electroencephalographic (EEG) data. Considering the EEG-specific topographical cortical regions as topographical modules (TMs), we find that the FC comprising positive correlations (G+) is modular. In contrast, networks of negative-correlations (G-) are anti-modular, with more connections between TMs than within them, and are associated with improved overall topological efficiency. These functional networks also show variability across frequency bands and brain states. In the low-frequency delta band, resting states exhibit higher modularity and anti-modularity than task states; in contrast, in the high-frequency Gamma band, modularity and anti-modularity are much higher during task states than in the resting state. The k-core analysis of all networks further reveals differences: G+ is more hierarchical and robust than G- across all states. Moreover, the task-state networks are always more hierarchical than the resting-state networks across all frequency bands. In the high-frequency gamma band, they are also significantly more robust than the resting-state networks. These networks also differ in the topology of their innermost core constituents: the innermost core regions of G+ are randomly connected and spatially localized, mostly in posterior brain regions across subjects, in the high-frequency gamma band. Whereas those in G- are spatially de-localized, cover the extreme anterior and extreme posterior brain regions, and remain anti-modular in all the frequency bands. Overall, our analysis reveals the presence of an anti-modular organization of functionally specialized TMs alongside their modular organization and points to task- and resting-state differences in their topologies.

neuroscience↗

bioRxiv2026

qPCR Guru, a free browser-based platform, strengthens microRNA analyses using full-curve Cq estimation

Quantitative PCR (qPCR) depends on reliable quantification cycle (Cq) estimation from amplification curves, which are not always well-behaved. We developed qPCR Guru to provide a complete analysis pipeline including data quality assessment, relative quantification, standard-curve diagnostics, and dual-method Cq evaluation. The latter compares the conventional instrument-derived threshold ("Reported") Cq versus the full-curve five-parameter logistic (5PL) second-derivative-maximum ("Fit") Cq and automatically flags curve-shape abnormalities and disagreement between the two estimates. On high-expressing targets (mRNA and microRNA), the two methods showed strong convergence, confirming general-purpose performance. On low-expressing targets, such as serum microRNA, baseline artifacts and biphasic amplification result in threshold miscalls that standard instrument analysis does not flag. Fit Cq restored replicate-concordant values where Reported Cq split the technical replicates by 17-20 cycles, recovered MIQE-compliant amplification efficiencies lost to biphasic miscalls (from 74% to 102% and 387% to 98%), and lowered within-group variability by 48% and 68% in feline and bovine samples, respectively. Together, these results demonstrate that full-curve estimation, with integrated curve-level diagnostics, strengthens qPCR analyses against threshold miscalls. ARTICLE HIGHLIGHTSO_LIqPCR Guru is a free, browser-based platform that provides a complete analysis pipeline and facilitates side-by-side comparisons of an instruments threshold (Reported) Cq and a full-curve (Fit) Cq, from the five-parameter logistic fitting with second-derivative-maximum (SDM/cpD2). C_LIO_LIFor every well the application automatically flags curve-shape abnormalities and disagreement between the two Cq estimates. C_LIO_LIOn clean, high-expressing mRNA and microRNA targets, the two estimators (Reported Cq and Fit Cq) were strongly concordant and produced equivalent relative quantification with comparable precision. C_LIO_LIIn low-expressing serum microRNA, baseline artifacts and biphasic amplification produced threshold Cq miscalls of up to [~]20 cycles and were detected by curve-shape flags and/or method disagreement. C_LIO_LIThe full-curve Cq estimate recovered replicate-concordant values, restored MIQE-compliant amplification efficiencies, and reduced within-group variability in serum microRNA. C_LI

Molecular Biology↗

bioRxiv2026

Simulation of cell-size systems at long timescales with flexible protein structures

Protein behavior inside cells is dominated by the crowded nature of the intracellular environment. Progress in structure determination of proteins and protein complexes, based on advances in Artificial Intelligence, provides an opportunity for structure-based modeling of cellular phenomena. Such modeling at the atomic resolution has been advanced by the traditional simulation techniques, e.g. molecular dynamics. A recently developed docking-based approach implements Markov Chain Monte Carlo sampling of intermolecular energy landscapes, offering several orders of magnitude faster simulation protocols. The approach allows addressing much longer trajectories of macromolecular systems in the crowded intracellular environment at atomic resolution. The sampling by design avoids low-probability (high-energy) states, which greatly accelerates the simulation process. A notable feature of this docking-based approach is the rigid body approximation of protein structures. The rigid-body approximation had been the primary direction in the protein docking field up until recent developments in deep learning. The rigid-body approach should be quite robust for the higher energy transient interactions that dominate the highly crowded cellular environment, as they likely involve relatively small conformational change. However, it is less applicable to the low-energy protein-protein complexes, especially those involving flexible regions. We addressed this problem by incorporating AlphaFold3 top models of the protein complexes in the mapping of the intermolecular energy landscape, as representative of the low-energy configurations of the protein assembly. By the nature of the AlphaFold predictions, these models involve appropriate conformational change between unbound and bound structures. These low-energy docking poses are combined with the rigid-body docking predictions that cover the multiplicity of the transient interactions. Such combination directly addresses the conformational flexibility of proteins upon binding along with the multiplicity of the transient protein encounters in the crowded cellular environment. SIGNIFICANCEProtein behavior inside cells is dominated by the crowded nature of intracellular environment. A recently developed approach allowed addressing long simulation trajectories of macromolecular systems in such environment at atomic resolution. A notable feature of this approach is the rigid body approximation in representation of the protein structures, which had been popular in the field up until the recent developments in artificial intelligence. However, such approximation is less applicable to stable protein-protein complexes, especially those involving flexible regions. We addressed this problem head-on by incorporating top deep learning-generated models of protein complexes. The new approach directly accounts for the flexibility of protein structures upon binding, along with the multiplicity of the transient protein encounters in the crowded cellular environment.

biophysics↗

bioRxiv2026

Region-specific molecular regulatory programs define epithelial identity, progenitor states, and mucus homeostasis in human distal airways

Small distal airways differ from proximal large airways in structure, airflow dynamics, and epithelial composition, and represent a central site of muco-obstructive lung disease pathogenesis. However, due in part to their inaccessibility, the molecular mechanisms that establish regional epithelial identity and govern mucociliary defense in distal airway epithelia remain poorly defined. Here, we integrate transcriptomic, secretomic, and chromatin accessibility analyses of matched primary human large and small airway epithelial cultures to define region-specific regulatory networks. We identify distal airway-specific transcriptional and chromatin programs required for maintaining epithelial identity and mucus homeostasis. Loss of NKX2-1 impairs distal airway secretory cell (DASC) differentiation and shifts mucus properties toward a disease-associated state. Lineage-resolved organoid assays identify an NKX2-1-high distal airway basal cell population with hybrid basal-secretory features as a selective progenitor for DASCs. Collectively, these findings establish a molecular framework for distal airway epithelial biology and define mechanisms regulating region-specific mucociliary host defense.

Cell Biology↗

bioRxiv2026

Heritable single-cell gene expression states shape functional variability in innate immune responses

Activation of innate immunity at the single-cell level is inherently heterogeneous, yet the mechanisms underlying this variability remain incompletely understood. Here, we integrate transcriptomics, high-content imaging and mathematical modelling to quantify transcriptional heritability within the evolutionarily conserved Toll-like receptor (TLR) system. RNA-seq-based fluctuation tests identified a subset of TLR4-dependent genes, including cytokines and immune effectors, that retain transcriptional heritability for more than 25 cell divisions in clonal macrophage populations. High-content microscopy confirmed gene-specific propagation of heritable states and revealed that environmental context shapes their persistence and expression. CD36, a scavenger receptor involved in bacterial recognition and lipid uptake, exhibited a stable, cell density-reinforced heritable state, whereas the inflammatory programme exemplified by IL1{beta} was transient, with heritability decaying upon clonal expansion. The interplay between heritable transcriptional states and population context generated emergent spatial organisation in high-density populations, with CD36-high cells forming discrete pockets and IL1{beta}-high cells enriched in surrounding regions. Functionally, CD36 expression determined clonal susceptibility to Listeria monocytogenes infection, linking transcriptional heritability to heterogeneous infection outcomes. Together, these findings identify transcriptional heritability as a key determinant of innate immune heterogeneity and demonstrate how heritable cellular states interact with population context to generate complex immune behaviours. Key point summaryO_LIMemorySeq and scRNA-seq identify long-term heritable gene expression states within TLR4-induced macrophage populations. C_LIO_LIHigh-content imaging across thousands of clonal populations reveals gene-specific dynamics of heritable states at the protein level. C_LIO_LIIl1{beta} and Cd36 define mutually exclusive heritable states that are differently regulated by population context and drive spatial organisation in cellular monolayers. C_LIO_LIHeritable CD36 protein expression shapes heterogeneous outcomes during Listeria monocytogenes infection. C_LI

immunology↗

bioRxiv2026

Developmental Continuity of Brain Network Core Organization in C. elegans

The brain is the most captivating chef doeuvre of nature. Naturally then, the mind wonders about the process that births such a fascinating organ. Neurodevelopment is a complex yet robust phenomenon that conceals answers to our questions in its intricacies. In an attempt to shed some light on this matter, we study the developing brain connectome of the nematode, C. elegans across the post-embryonic phase. A tiny organism with only around 200 neurons comprising its brain and yet a diverse array of behaviors to display, it makes for a great model. Starting with most of its head neurons already present at hatching, the worm brain accumulates numerous more synaptic connections increasing the edge density. It maintains a weak connectivity throughout thereby, balancing global communication as well as hierarchy. At the mesoscopic level, we find that the core has a conserved backbone of persistent neurons along with a dynamic component formed of transient/recurring neurons. Moreover, the connectome has a rich club organization since the early stage which selectively strengthens indicating progressively denser connectivity among the integrators due to the previously reported asymmetric synapse addition. This asymmetry also shows up in the preservation of input hubs across development and the progressively more centralized organization of the in-degree k-core. Our work provides a new perspective into the neurodevelopment of the brain that may facilitate our understanding of its functioning.

neuroscience↗

bioRxiv2026

FMR1 gene therapy restores activity-driven inhibition and prevents audiogenic seizures in Fmr1-/y mice

Fragile X syndrome (FXS) is a neurodevelopmental disorder associated with auditory hypersensitivity, circuit hyperexcitability, and seizures. Whether re-expression of the FMR1 gene and encoded Fragile X Messenger Ribonucleoprotein (FMRP) can restore sensory circuit dysfunction remains unclear. Here, we show that a viral AAV-FMR1 vector rescues audiogenic seizures in the Fmr1-/y mouse model after both neonatal and adult delivery, indicating that auditory circuit dysfunction remains reversible. Local re-expression in the inferior colliculus (IC) is sufficient to reduce seizure susceptibility, identifying this region as a key site of FMRP-dependent circuit regulation. In the IC, Translating Ribosome Affinity Purification and RNA-seq (TRAP-seq) profiling reveals impaired induction of sound-evoked translation programs in Fmr1-/y neurons, including those regulated by transcription factor Npas4. AAV-FMR1 restores a WT-like molecular response and normalizes unbalanced sound-evoked activation of VGLUT2+ excitatory neurons over VGAT+ inhibitory neurons in Fmr1-/y IC. Together, these findings indicate altered translation of Npas4 in response to sound impairs recruitment of inhibition in the Fmr1-/y IC, and this can be reversed with AAV-FMR1 administration. Moreover, the rescue of seizures after adult administration of AAV-FMR1 supports a gene therapy approach for FXS. HighlightsO_LITRAP-seq reveals impaired activity-driven translation in Fmr1-/y inferior colliculus (IC) C_LIO_LIDeficient Npas4 induction reduces evoked inhibition in Fmr1-/y IC C_LIO_LIAAV9-FMR1 gene therapy normalizes translation and excitatory/inhibitory balance in Fmr1-/y IC C_LIO_LIAAV9-FMR1 gene therapy prevents audiogenic seizures in Fmr1-/y mice when administered neonatally or in adulthood C_LI

neuroscience↗

bioRxiv2026

Amino acid repeat mosaics shape protein functional landscape

How combinations of functional units such as repeats and motifs within disordered regions influence protein functions remains poorly understood. Here, we investigate how co-occurring amino acid homorepeats of different types within the same protein (HR-mosaics) shape molecular outcomes. Using a novel evolutionary co-occurrence metric, human HR-mosaics were classified as segregated (each HR independently influences distinct outcomes), concerted-disjunct (both HRs jointly shape protein functionality without modulating each other), or concerted-conjunct (both HRs affect functionality through mutual influence). Molecular studies of naturally occurring polyGly-polyPro mosaic in DDX17 and chimeric mosaics of polyAla and polyHis demonstrate that HR-mosaics influence protein abundance, localisation, mobility and interaction landscape. Segregated HR-mosaics expand functional space, concerted-disjunct additionally confer biological context, and concerted-conjunct mosaics further enable rheostatic regulation between HRs. These design principles underscore amino acid type, length and relative positioning of HRs as central architects of HR-mosaic functionality, with implications in protein design and engineering and understanding repeat-associated pathologies.

Systems Biology↗

bioRxiv2026

Bacterial metabolism of synthetic steroids across ecosystems reveals diverse biotransformation products, reactions, and enzymes

Biotransformation of synthetic steroid drugs by gut and environmental bacteria affects their activity within the host, persistence, and environmental fate of these widely used pharmaceuticals. However, the spectrum of bacterial transformation of synthetic steroids, the enzymes involved, and the role of cooperative microbial metabolism in these processes remain poorly understood. We systematically mapped the biotransformation of 22 steroids, comprising 20 synthetic and 2 natural compounds and spanning clinically used estrogens, progestogens, corticosteroids, and prodrugs, across 12 bacterial species (8 intestinal and 4 environmental), yielding 264 steroid-bacteria combinations. We identified 97 unique biotransformation products and found that the tested bacteria catalyze diverse reactions including ester hydrolysis, oxidation-reduction chemistry, and steroid side-chain cleavage. Notably, the environmental bacterium Sphingobium herbicidovorans catalyzed desmolase-like steroid side-chain cleavage under aerobic conditions, a transformation previously reported solely for anaerobic bacteria. Combining sequence homology searches, a genetic gain-of-function screen, and expression proteomics we identified six steroid-transforming enzymes in S. herbicidovorans. We further demonstrate that distinct bacterial species cooperatively metabolize synthetic steroids through cross-feeding, enabling sequential activation and metabolism of corticosteroids across microbial communities. Together, our findings uncover previously unrecognized bacterial enzymes and community-level interactions involved in synthetic steroid metabolism. By directly linking steroid metabolites, enzymes, and microbial community interactions, this study provides molecular-level mechanistic insights into microbial steroid biotransformation. Such insights are essential for ultimately predicting metabolic interactions within and across microbial communities, as well as their interactions with the host and the environment.

microbiology↗

bioRxiv2026

Neural decoding of speech using deep neural ensembles

Speech brain-computer interfaces (BCIs) can restore rapid communication to people with paralysis, but decoding errors still limit performance. In recent brain-to-text decoding competitions, deep ensemble methods, which combine predictions from multiple independently trained decoders, have delivered striking accuracy improvements and account for the largest gains over baseline approaches. However, these methods have not previously been tested in real-time, require substantial computational resources, and their performance under various clinically relevant constraints remains poorly understood. Here, we present the first closed-loop test of deep ensembles in a participant with bilateral intracortical microelectrode arrays, demonstrating a reduction in word error rate from 33.7% to 26.0% on a large-vocabulary task. Using additional data from three participants, we then assess how these gains depend on baseline error rate, training dataset size, and ensemble size, including the resource-accuracy tradeoffs most relevant for real-world deployment. Finally, we introduce a computationally efficient pseudoensembling approach based on test-time augmentation that improves decoding accuracy while requiring only a single base decoder, greatly reducing the computational burden of ensembling. Together, these results show that the benefits of deep ensembling can be realized in real time and under practical resource constraints, bringing speech BCIs closer to broader clinical adoption.

neuroscience↗

bioRxiv2026

A genetically encoded biosensor reveals heterogenous cAMP dynamics coordinating growth and resuscitation in Mycobacterium tuberculosis

3',5'- Cyclic adenosine monophosphate (cAMP) serves as a global regulatory hub in Mycobacterium tuberculosis (Mtb), which dedicates approximately 1% of its genome to the synthesis, degradation and sensing of this secondary messenger. Despite cAMPs central role in Mtb virulence and metabolic adaptation, we currently do not have any tools to monitor cAMP dynamics in real time at single-cell resolution. Here, we report the development and validation of rmgCarvi, a genetically encoded biosensor adapted from eukaryotic systems for use in mycobacteria. By utilizing cAMP-responsive cpGFP and constitutive mCherry fluorescence, rmgCarvi provides a high-fidelity ratiometric readout of intracellular cAMP levels, revealing that carbon source and growth state act as primary determinants of cAMP dynamics. Real-time single-cell imaging using rmgCarvi demonstrated robust correlations between cAMP and growth rate, with pronounced cAMP spikes coinciding with cell division. In starved Mtb populations, a bimodal distribution in cAMP levels dictates a distinct hysteresis during regrowth upon nutrient supplementation. Low-cAMP subpopulations are associated with strong induction in cAMP levels and a longer lag before initiation of growth, while high cAMP cells do not exhibit significant induction in cAMP and have recovery times uncorrelated with prior cAMP levels. Finally, use of rmgCarvi in infection studies captured host species-specific responses, including a strong cAMP induction upon internalization followed by a decline in murine, but not human, macrophages. By resolving cAMP dynamics with unprecedented cellular and temporal resolution, rmgCarvi provides a framework for understanding how Mtb tunes this crucial metabolite to navigate the complex microenvironments in the host. Significance StatementWe report the evaluation of a ratiometric genetic biosensor, rmgCarvi, for real-time single-cell monitoring of cAMP dynamics in mycobacteria. This biosensor reveals how Mycobacterium tuberculosis precisely tunes cAMP within a narrow range ("Goldilocks zone") to coordinate metabolic adaptation, and growth across complex nutrient environments. In nutrient-starved populations, we identify a bimodal cAMP distribution that functions as a phenotypic memory, dictating resuscitation kinetics upon nutrient restoration. As the first cAMP biosensor validated in any mycobacterial species, rmgCarvi facilitates the discovery of host-specific signaling dynamics during infection and provides a platform for screening of compounds that perturb cAMP homeostasis. This study provides a framework to explore how Mtb utilizes this secondary messenger molecule to survive and adapt within complex host environments.

microbiology↗

bioRxiv2026

Two Genomes, One Metabolome: Mitonuclear Incompatibility Remodels Developmental Metabolism and Fitness in Drosophila

Mitochondrial-nuclear (hereafter mitonuclear) genetic variation can alter cellular bioenergetics and metabolism via jointly encoding the subunits of oxidative phosphorylation (OXPHOS) system. Here we tested whether genetic variation in biochemical and bioenergetic phenotype can scale up across higher levels of biological organization to affect organismal development. We used a panel of (mtDNA); nDNA genotypes created by asymmetric substitution of divergent mtDNA between Drosophila melanogaster and its sister species D. simulans. Two genotypes - (ore); OreR and (ore); Aut - carry coevolved mitonuclear genomes from D. melanogaster, whereas the other two - (w501); Aut and (w501); OreR - harbor D. simulans mitochondrial DNA introgressed onto D. melanogaster nuclear backgrounds. We utilized untargeted metabolomics to track down the comprehensive biochemical footprint of the joint genomic architecture in the metabolome of these genotypes. We show that in (w501); OreR larvae with mitonuclear genome incompatibility there is extensive and coordinated metabolic remodeling of carbon, nitrogen, and redox balance, characterized by increased glycolysis, limited TCA cycle, amino acid scarcity, reduced nucleic acid balance, lipid remodeling, with compensatory activation of antioxidant pathways and possible epigenetic modulation by metabolites. This biochemical rewiring puts a physiological constraint that prioritizes maintenance metabolism over larval growth, leading to reduced body size, slow locomotion and delayed development, despite a compensatory increase in feeding in these larvae. These findings underscore the context dependency of mitonuclear interactions, which can uniquely scale up to influence organismal energy budget trade-offs, fitness and life-history evolution.

physiology↗

bioRxiv2026

Scalable multimodal mapping of macrophage regulatory architecture by integrating optical and transcriptomic pooled screens

Understanding how genetic perturbations reshape cellular states requires measuring diverse phenotypic modalities at scale. Here, we present PerturbPair, a platform that combines parallel Perturb-Seq and optical pooled screening (OPS) in primary mouse bone marrow-derived macrophages stimulated with lipopolysaccharide. Profiling over 334,000 single-cell transcriptomes across [~]1,000 gene perturbations and 7.8 million imaging-phenotyped cells across [~]3,000 gene perturbations, we reveal high concordance between transcriptomic and optical perturbation signatures. While RNA and imaging phenotypes were broadly concordant, OPS demonstrated superior sensitivity for weak-effect perturbations owing to greater cell throughput, and captured post-transcriptional regulatory events--such as protein trapping and mTOR-dependent phosphorylation--that left no detectable transcriptional footprint. To exploit cross-modal relationships, we developed EB-MoCAVI, an empirical Bayesian variational inference framework that both imputes RNA profiles for perturbations measured only by imaging--effectively tripling our Perturb-Seq dataset in silico--and denoises transcriptomic estimates for perturbations with sparse cell coverage by leveraging matched imaging data as a regularizer. A secondary screen validated the imputed transcriptional profiles and corroborated the cytosolic iron-sulfur assembly pathway as a candidate restraint on macrophage interferon tone. Integrating measured and imputed profiles with rare-variant burden statistics from the UK Biobank identified disease-specific macrophage gene programs and causal regulatory nodes for monocyte counts, type 2 diabetes, and inflammatory bowel disease. PerturbPair establishes a generalizable framework for multimodal perturbation atlases, pointing toward quantitative, causally-informed cross-modal models of cellular behavior.

genomics↗

bioRxiv2026

DMPKformer: An Interpretable Multimodal Deep Learning Framework for Reliable ADMET Property Prediction

Accurate prediction of absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties remains a critical challenge in drug discovery. Traditional single modality approaches often fail to capture the complex, multi-scale relationships governing molecular behavior across physicochemical, structural, and pharmacokinetic dimensions. In this work, we propose a multi-modal deep learning framework that integrates complementary molecular representations, MACCS fingerprints, molecular graphs, and physicochemical descriptors to achieve robust ADMET property prediction. Each modality is modeled using a specialized neural subnetwork tailored to its structural characteristics: a self-attention-based Transformer encoder for MACCS fingerprints, a Graph Attention Network (GAT) for molecular graph representations, and a tanh-activated multilayer perceptron for RDKit-, PaDEL-, and Mordred-derived descriptors. Each modality is independently trained for binary classification, and latent embeddings extracted from internal layers serve as transferable molecular representations. These embeddings are subsequently fused and fine-tuned via a tanh-activated dense network and shared prediction head to form a unified ADMET predictor. The proposed framework achieves competitive performance across multiple TDC ADMET benchmarks while providing enhanced interpretability through modality-specific attention mechanisms. In addition, the incorporation of latent-space out-of-distribution (OOD) confidence estimation enables identification of high-confidence operating regions, improving the reliability and practical applicability of the framework for molecular property prediction in drug discovery workflows.

bioinformatics↗