Biology subjects

Pang, K.

Publications and source records attributed to Pang, K..

15 recordsLinked to original sources

bioRxiv2026

Perturb-ME: Scalable mechanism discovery from phenotype-enriched genome-wide screens

Perturb-seq enables pooled genetic screens with rich single-cell profiling readouts, but genome-scale profiling remains costly and may not be associated with other established functional characteristics. Moreover, as screens grow in size and complexity, interpreting the resulting data comprehensively is challenging and slow. Here, we introduce Perturb-seq with Marker Enrichment (Perturb-ME), which combines genome-scale CRISPR screening, phenotype-based enrichment and multimodal single-cell profiling. Applied to MHC-I cell surface protein expression in melanoma, Perturb-ME profiled HLA-low and HLA-high cells with matched RNA, surface-protein and guide measurements. A regulatory model with 221 impactful regulators affecting 1,998 responsive genes recovered seven coherent co-functional regulatory modules governing nine gene programs, including the canonical IFN{gamma}-MHC-I axis regulating an antigen-presentation and interferon-response program. Agentic interpretation of the entire model with an AI co-scientist linked additional modules to trafficking, proteostasis and chromatin regulation. Perturb-ME, along with agentic interpretation, provide a scalable framework for comprehensive functional discovery from phenotype-enriched genetic screens.

cell biology↗

bioRxiv2026

Dissociation between impaired explicit spatial remapping and preserved implicit neural dynamics in Alzheimers disease

The hippocampus supports spatial memory by dynamically integrating external sensory inputs with intrinsic neural circuit dynamics during novel experience. In Alzheimers disease (AD), despite impaired hippocampal spatial remapping, spatial learning and memory abilities can remain partially preserved, a phenomenon consistent with cognitive resilience (Gomez-Isla and Frosch 2022, Jia, Xu et al. 2025). However, the hippocampal ensemble coding patterns associated with these preserved learning and memory abilities remain remains unclear. We hypothesize that intrinsic temporal structures of neuronal firing continue to facilitate the encoding of new spatial information. Using the AppNL-G-F rat model, we longitudinally tracked hippocampal CA1 activity during a familiar-novel context alternating task. We found a dissociation between impaired explicit spatial coding and preserved implicit temporal coding in the AD hippocampal network. Explicit spatial coding was impaired, as place cells showed weak discrimination between distinct contexts and failed to improve with learning. In contrast, implicit temporal coding exhibited learning-dependent refinement, with cofiring dynamic becoming increasingly context-specific across long-term experience. Further analysis suggested that the enhancement of implicit cofiring may be associated with the increased consistency of neural ensemble reactivation during sharp wave ripples in awake rest. Taken together, these findings reveal an explicit-implicit dissociation in the AD hippocampal network, suggesting that the learning-dependent refinement of implicit temporal coding may support preserved learning capacity despite impaired spatial remapping.

neuroscience↗

bioRxiv2026

Non-contact direct sensing of material properties of biomolecular condensate using Scanning Ionic Conductance Microscopy

Biomolecular condensates are important regulators of cellular compartmentalization and biochemical processes. Understanding their material properties is critical to elucidate how they control molecular organization and dynamics within cells. However, quantitatively probing these properties remains challenging due to the wide range of length scales, concentrations, and timescales over which condensates operate, as well as the limited force ranges accessible to current nanoscale mechanical mapping methods. We explored the use of a non-contact 3D imaging tool Scanning Ion Conductance Microscopy (SICM) for stiffness measurements of liquid-liquid phase-separated biomolecular condensates. We focus on the Dhh1 protein, which is a regulator of cytoplasmic processing bodies (PBs) membrane-less cytoplasmic condensates that control the storage and degradation of untranslated mRNA. In our study, we investigate the properties of mCherry2- or His-mCherry2-tagged full-length Dhh1 and N- or C-terminus tail-deletion constructs, as well as the catalytically inactive mutant DQAD, under different pH and incubation times. We mapped both spatial and temporal changes in the material properties of the condensates, highlighting the capabilities of the instrument. We found that the removal of either of the two tails led to an increase in condensate stiffness upon shifting the pH from a stress-associated cellular environment (pH 6.5) to physiological conditions (pH 7.5). Additionally, the choice of protein tags led to vastly different results depending on the pH where mCherry2-Dhh1 exhibited a stiffening going from pH 6.0 to 6.5 while the double-tagged His-mCherry2 did not. Our measurements are verified and corroborated by established techniques such as optical tweezer-based fusion assays and fluorescence recovery after photobleaching (FRAP). Furthermore, we were able to track the same biomolecular condensate sample for up to 20 days getting insights on the ageing and evolution of the condensates. Overall, our study demonstrates the applicability of SICM for direct measurement of the material properties of biomolecular condensate.

biophysics↗

bioRxiv2026

Hypothalamic MCH neurons links tau pathology to sleep disruption

Sleep disruption is an early and pervasive feature of Alzheimers disease (AD), yet the circuit mechanisms linking tau pathology to sleep-wake dysfunction remain unresolved. Here, we identify melanin-concentrating hormone (MCH) neurons in the lateral hypothalamus (LH) as a critical node disrupted in tauopathy. Longitudinal EEG/EMG recordings in PS19 mice reveal progressive impairments in sleep architecture and homeostasis. Histological analyses revealed significant degeneration of both MCH neurons and the neighboring hypocretin (Hcrt) neuronal population in the LH at late stages of tauopathy. In vivo fiber photometry recordings demonstrated a selective functional impairment of MCH neurons, characterized by reduced activity during REM sleep, whereas Hcrt neuronal activity remained largely preserved. In addition, cell-autonomous expression of mutant tau in MCH neurons recapitulates sleep disruption, establishing a causal role. Despite tauopathy-induced neuronal loss and reduced endogenous activity, optogenetic and chemogenetic activation show that MCH neurons retain functional capacity, and their activation restores sleep in aged PS19 mice. Together, these findings define a circuit mechanism linking tau pathology to sleep and identify MCH neurons as a tractable therapeutic target.

neuroscience↗

bioRxiv2025

PULSAR: a Foundation Model for Multi-scale and Multicellular Biology

Biology emerges from interactions across physical scales, where molecular interactions drive cellular states, which in turn orchestrate multicellular tissue functions that collectively define health and disease. However, current computational models are often constrained to single scales in isolation, failing to integrate the biology that emerges from lower to higher levels [1]. Here we present PULSAR (Patient Understanding Leveraging Single-cell universAl Representation), a multi-scale and multicellular foundation model architecture that explicitly enables information flow from genes to cells to multicellular systems. Applied to the human peripheral immune system, PULSAR extracts a unified donor representation that supports rapid disease classification, biomarker prediction, and forecasting of future clinical events, such as Rheumatoid arthritis onset. As a generative model, PULSAR enables the simulation of cytokine perturbation response across physical resolutions, while its interpretability reveals the key cell types driving disease. Overall, PULSAR opens new avenues for precision medicine by enabling computational reasoning that connects molecular biology to clinical phenotypes.

bioinformatics↗

bioRxiv2025

LUMI-lab: a Foundation Model-Driven Autonomous Platform Enabling Discovery of New Ionizable Lipid Designs for mRNA Delivery

The complexity of molecular discovery requires autonomous systems that efficiently explore vast and uncharted chemical spaces. While integrating artificial intelligence (AI) with robotic automation has accelerated discovery, its application remains constrained in fields with scarce historical data. One such challenge is the design of lipid nanoparticles (LNPs) for mRNA delivery, which has relied on expert-driven design and is hindered by limited datasets. Here, we introduce LUMI-lab, a self-driving lab (SDL) system that enables efficient learning with minimal wet-lab data by integrating a molecular foundation model with an automated active-learning experimental workflow. Through ten iterative cycles, LUMI-lab synthesized and evaluated over 1,700 LNPs, identifying ionizable lipids with superior mRNA transfection potency in human bronchial cells compared to clinically approved benchmarks. Unexpectedly, it autonomously uncovered brominated lipid tails as a novel feature enhancing mRNA delivery. In vivo validation further confirmed that inhalation of LNPs containing the top-performing lipid, LUMI-6, achieved 20.3% gene editing efficacy in lung epithelial cells in murine models, surpassing the highest efficiency reported for inhaled LNP-mediated CRISPR-Cas9 delivery in mice to our knowledge. These findings demonstrate LUMI-lab as a powerful, data-efficient platform for advancing mRNA delivery, highlighting the potential of AI-driven autonomous systems to accelerate innovation in material science and therapeutic discovery.

biochemistry↗

bioRxiv2025

The xenacoelomorph gonopore is homologous to the bilaterian anus

The bilaterian through gut with an anal opening is a key invention in animals, since it facilitates effective food processing, which allows animals to grow to a larger body size. However, because non-bilaterian animals lack a through gut, the evolution of anus is still debated. The formation of bilaterian hindgut is governed by the spatial expression of several transcription factors (e.g. Caudal and Brachyury) under the control of Wnt signaling. This conserved pattern has been used to support the homology of the anus of protostomes (insects, snails) and deuterostomes (sea urchins, humans). Here we show, that these bilaterian "hindgut" marker genes are expressed around the male gonopore of several xenacoelomorphs, which have a blind gut without an anal opening. These findings suggest a deep evolutionary relationship between the xenacoelomorph male gonopore and the bilaterian anus. Since xenacoelomorphs are the potential sister group to all remaining Bilateria, our results suggest that the bilaterian anus evolved from a male gonopore that came in contact with the digestive endoderm to form the posterior opening.

developmental biology↗

bioRxiv2023

Potential clinical value and influence of conductivity in conductive cardiac patches for reducing post-MI arrythmia risks

Conductive cardiac patches can help to restore electric signal conduction of the diseased myocardium after myocardial infarction (MI). However, none of the conductive cardiac patches reported in literature has entered clinical trials. Bench-to-bedside translation of conductive patches has long been hindered by the lack of knowledge of the optimal patch conductivity and deep understanding of the potential clinical benefits and risks in patients. Here, we first prepared conductive cardiac patches with conductivities covering 5 orders of magnitude (10-3-101 S/cm). Disagreeing with the mainstream opinion that patch conductivity close to native myocardium (10-3-10-2 S/cm) is most favorable, our results showed that patches with conductivity two orders of magnitudes higher than native myocardium (10-1-100 S/cm) are most effective in restoring cardiac conduction and lowering inducibility quotient. Conduction velocity (CV) is the essence of the observed results. Rat experiments showed that the low-conductivity patch could increase the CV of infarcted myocardium, but did not fully compensate the mismatch in the CVs of infarct and healthy myocardium. Moderate-conductivity patches could increase myocardial CV to the same level of healthy myocardium, while high-conductivity patches further increased myocardial CV, causing a reversed mismatch. The relationship between patch conductivity and improved CVs in myocardium can be explained by monodomain model theory. Based on the theory, 3D finite element simulation of a MI patient heart predicted that a suitable, patch-improved myocardial CV could reduce the number of reentrants, and stabilize the remaining reentry circuits in the myocardium of the MI patient, which indicated its clinical value.

bioengineering↗

bioRxiv2023

Deciphering 3'UTR mediated gene regulation using interpretable deep representation learning

The 3untranslated regions (3UTRs) of messenger RNAs contain many important cis-regulatory elements that are under functional and evolutionary constraints. We hypothesize that these constraints are similar to grammars and syntaxes in human languages and can be modeled by advanced natural language models such as Transformers, which has been very effective in modeling protein sequence and structures. Here we describe 3UTRBERT, which implements an attention-based language model, i.e., Bidirectional Encoder Representations from Transformers (BERT). 3UTRBERT was pre-trained on aggregated 3UTR sequences of human mRNAs in a task-agnostic manner; the pre-trained model was then fine-tuned for specific downstream tasks such as predicting RBP binding sites, m6A RNA modification sites, and predicting RNA sub-cellular localizations. Benchmark results showed that 3UTRBERT generally outperformed other contemporary methods in each of these tasks. We also showed that the self-attention mechanism within 3UTRBERT allows direct visualization of the semantic relationship between sequence elements.

bioinformatics↗

bioRxiv2023

Tumor microenvironment exploration and therapeutic target identification in bladder cancer based on Cuproptosis signature and chemotherapy-resistant model: results from bioinformatics and experimental validation

BackgroundThe discovery of cuproptosis provides a new way to make full use of the pathophysiological effects of copper for anticancer therapy and could help identify a therapeutic target in bladder cancer. MethodsIn this study, 411 BLCA tumor samples from the Cancer Genome Atlas (TCGA) cohort were obtained. Differentially expressed genes (DEGs) between chemotherapy-sensitive and chemotherapy-resistant mouse bladder cancers were also obtained. Sixteen genes were defined as cuproptosis-related genes (CRGs), and the cutoff score was calculated based on LASSO Cox regression. CCK-8 and Transwell assays were used to detect the migration and proliferation of 5637 and T24 cells, respectively. Liarozole dihydrochloride (L-D), a mild P450, inhibits the expression of CYP26B1. Multiple immunohistochemistry analyses were used to explore the association between the immune microenvironment and CRGs. ResultsA higher Cuproptosis score was significantly associated with worse overall survival in BLCA (p<0.0001, HR=2.32). Single-cell transcriptional data were used to assess the function of CYP26B1 and CYP26B1 may be negatively associated with DNA damage and repair. In vitro experiments indicated that overexpression of CYP26B1 enhanced the migration and proliferation of BLCA cells, while L-D inhibited the migration and proliferation of BLCA cells. A higher expression level of Dihydrolipoamide S-Succinyltransferase (DLST) serves could as a risk factor in patients treated with atezolizumab and higher expression of DLST suggest an immunosuppressive microenvironment. ConclusionsCYP26B1 may be a therapeutic target for bladder cancer, and the higher expression of DLST may suggest an immunosuppressive microenvironment in BLCA. Key MessagesCuproptosis may be associated with the chemotherapy-resistance of bladder cancer and our findings may provide new ideas for the treatment of bladder cancer.

genomics↗

bioRxiv2022

A cross-species proteomic map of synapse development reveals neoteny during human postsynaptic density maturation

The molecular mechanisms and evolutionary changes accompanying synapse development are still poorly understood. Here, we generated a cross-species proteomic map of synapse development in the human, macaque, and mouse neocortex. By tracking the changes of >1,000 postsynaptic density (PSD) proteins from midgestation to young adulthood, we found that PSD maturation in humans separates into three major phases that are dominated by distinct pathways. Cross-species comparisons reveal that the human PSD matures about two to three times slower than other species and contains higher levels of Rho guanine nucleotide exchange factors (RhoGEFs) in the perinatal period. Enhancement of the RhoGEF signaling in human neurons delays the morphological maturation of dendritic spines and the functional maturation of synapses, potentially contributing to the neotenic traits of human brain development. In addition, PSD proteins can be divided into four modules that exert stage- and cell type-specific functions, possibly explaining their differential associations with cognitive functions and diseases. Together, our proteomic map of synapse development provides a blueprint for studying the molecular basis and evolutionary changes of synapse maturation.

neuroscience↗

bioRxiv2020

Co-expression enrichment analysis at the single-cell level reveals convergent defects in neural progenitor cells and their cell-type transitions in neurodevelopmental disorders

Recent large-scale sequencing studies have identified a great number of genes whose disruptions cause neurodevelopmental disorders (NDDs). However, cell-type-specific functions of NDD genes and their contributions to NDD pathology are unclear. Here, we integrated NDD genetics with single-cell RNA sequencing data to identify cell-type and temporal convergence of genes involved in different NDDs. By assessing the co-expression enrichment pattern of various NDD gene sets, we identified mid-fetal cortical neural progenitor cell development--more specifically, ventricular radial glia-to-intermediate progenitor cell transition at gestational week 10--as a key convergent point in autism spectrum disorder (ASD) and epilepsy. Integrated gene ontology-based analyses further revealed that ASD genes function as upstream regulators to activate neural differentiation and inhibit cell cycle during the transition, whereas epilepsy genes function as downstream effectors in the same processes, offering a potential explanation for the high comorbidity rate of the two disorders. Together, our study provides a framework for investigating the cell-type-specific pathophysiology of NDDs.

genetics↗

bioRxiv2019

SIDT2 RNA transporter promotes lung and gastrointestinal tumor development

RNautophagy is a newly-described type of selective autophagy whereby cellular RNAs are transported into lysosomes for degradation. This process involves the transmembrane protein SIDT2, which transports double-stranded RNA (dsRNA) across the endolysosomal membrane. We previously demonstrated that SIDT2 is a transcriptional target of p53, but its role in tumorigenesis - if any - is unclear. Unexpectedly, we show here that Sidt2-/- mice with concurrent oncogenic KrasG12D activation develop significantly fewer tumors than littermate controls in a mouse model of lung adenocarcinoma (LUAD). Consistent with this observation, loss of SIDT2 also leads to enhanced survival and delayed tumor development in an Apcmin/+ mouse model of intestinal cancer. Within the intestine, Apcmin/+;Sidt2-/- mice display accumulation of dsRNA in association with increased phosphorylation of eIF2 and JNK as well as elevated rates of apoptosis. Taken together, our data demonstrate a role for SIDT2 - and by extension RNautophagy - in promoting tumor development.

cancer biology↗

bioRxiv2019

An high dose of a conjugated linoleic acid increases fatty liver and insulin resistance in lactating mice

This study aimed to evaluate the effect of a high dose of conjugated linoleic acid (CLA) on lactating mice. In experiment one, KUNMING mice were separated into four groups (n = 6 per group); the control (CON) group received 3.0% linoleic acid oil (LA), the L-CLA group received 1.0% LA and 2.0% CLA mixture, the M-CLA group received 2.0% LA and 1.0% CLA mixture, and the H-CLA group received 3.0% CLA mixture. Feeding proceeded from day 4 to day 10 during lactation. In experiment two, the CON group received 2.0% LA and the H-CLA group received 2.0% CLA. Blood parameters were analysed for all groups, and insulin tolerance tests (ITTs) were conducted. CLA treatment did not affect the dam weight, but it significantly decreased the food intake of dams. Furthermore, CLA decreased the weight of pups, which was attributed to lower milk fat. H-CLA group mice displayed increased liver weight and liver triglyceride content, as well as a higher TG content and {gamma}-GT activity in blood. Moreover, a high dose of CLA resulted in insulin resistance, possibly affecting the RBC and HCB of blood. In conclusion, lactating mice receiving a high dose of CLA led to fatty liver, insulin resistance, and impaired lactation performance.

physiology↗

bioRxiv2019

Anhedonia following mild traumatic brain injury in rats: A behavioral economic analysis of positive and negative reinforcement

Psychiatric disorders affect nearly 50% of individuals who have experienced a traumatic brain injury (TBI). Anhedonia is a major symptom of numerous psychiatric disorders and is a diagnostic criterion for depression. Recently, anhedonia has been divided into consummatory, motivational and decisional components, all of which may be affected differently in disease. Although anhedonia is typically assessed using positive reinforcement, the importance of stress in psychopathology suggests the study of negative reinforcement (removal or avoidance of aversive events) may be equally important. The present study investigated positive and negative reinforcement following a rat model of mild TBI (mTBI) using lateral fluid percussion. Hedonic value of and motivation for reinforcement was determined by behavioral economic analyses. Following mTBI, the hedonic value of avoiding foot shock was reduced. In contrast, the hedonic value of escaping foot shock or obtaining a sucrose pellet was not altered by mTBI. Moreover, neither motivation to avoid or escape foot shock nor motivation to acquire sucrose was modified by mTBI. Our results suggest that individuals experiencing mTBI may experience more stress because of poor proactive control of stress resulting from reduced hedonic value of avoiding aversive events.

animal behavior and cognition↗