Biology subjects

Ye, B.

Publications and source records attributed to Ye, B..

At least 19 recordsLinked to original sources

bioRxiv2026

LabGrymace: Automated Analysis of Mouse Grimace for Quantitative Assessment of Pain Dynamics

Accurate assessment of pain in animal models is essential for understanding pain mechanisms, developing analgesics, and ensuring animal welfare. The Mouse Grimace Scale (MGS) provides a sensitive, non-invasive measure of spontaneous pain by quantifying pain-related facial expressions, but its utility is limited by labor-intensive manual scoring, observer variability, and reliance on static images that fail to capture the temporal dynamics of facial behavior. Existing automated approaches improve throughput but typically rely on highly standardized imaging conditions, selected viewing angles, and static facial appearance, while providing limited temporal resolution and little insight into the relative contributions of individual facial action units. Here, we introduce LabGrymace, an open-source, artificial intelligence-powered framework for automated, frame-by-frame analysis of pain-related facial dynamics in freely moving mice. Built on the LabGym behavioral analysis platform, LabGrymace uses deep-learning-based facial feature detection and tracking to quantify ear, eye, and nose movements continuously from video recordings. To generate a quantitative pain metric and facilitate reproducibility, we calibrated facial dynamics against graded chemogenetic activation of nociceptors and identified the kinematic features most strongly associated with pain intensity. These features were integrated into a weighted composite pain score that reflects the differential contributions of individual facial action units. LabGrymace accurately classified pain-related facial actions and generated continuous pain scores without manual frame selection or restrictive recording conditions. The resulting pain scale exhibited dose-dependent responses generalized across distinct pain modalities, including visceral pain induced by MgSO and somatic pain induced by capsaicin. By combining automated facial-feature analysis with quantitative temporal modeling, LabGrymace provides an objective, scalable, interpretable, and flexible tool for assessing spontaneous pain in laboratory mice.

neuroscience↗

bioRxiv2026

G-SPRI: A Structure-Centric Graph Model for Comprehensive Prediction of Cancer Driver Events from Missense Mutations

In silico approaches for predicting the functional impact of missense mutations are critical for interpreting personal genomes and identifying disease-related biomarkers. Existing methods largely rely on sequence-based information or intuitive structural features, but often overlook the complex biophysical patterns encoded in protein 3D structures. Here, we present G-SPRI, a multilevel framework built on a novel alpha-shape protein graph that accurately captures residue connectivity from atomic-resolution geometry and enables precise message passing around mutation sites. Using this graph representation, G-SPRI integrates wild-type structural properties and mutation-specific perturbation signals derived from the Protein Data Bank (PDB) universe to support graph-based learning for distinguishing pathogenic from benign missense variants. G-SPRI performs strongly across multiple key tasks. On the binary prediction benchmark, G-SPRI delivers improved pathogenicity prediction for individual mutations. By integrating mutation recurrence across the pan-cancer cohort, G-SPRI recovers more known cancer driver genes than state-of-the-art methods from more than 2.3 million mutations. Furthermore, by jointly quantifying site-specific pathogenicity and co-clustering influence within higher-order structural organization units, G-SPRI provides comprehensive evidence for pinpointing likely driver mutations and structurally susceptible regions within disease genes.

bioinformatics↗

bioRxiv2026

PRMT5-Mediated Arginine Methylation of HSP90AA1 Drives Esophageal Squamous Cell Carcinoma Progression

Dysregulated HSP90AA1 chaperone activity is a hallmark of multiple human cancers, yet its post-translational modifications in esophageal squamous cell carcinoma (ESCC) are poorly defined. Here, we identify a PRMT5-driven modification of HSP90AA1, symmetric dimethylation at arginine 182 (R182), as a pivotal switch that fuels ESCC progression. HSP90AA1 physically associates with PRMT5, and genetic or pharmacologic PRMT5 blockade diminishes HSP90AA1-R182 methylation and downstream oncogenic signaling. Functionally, HSP90AA1 loss suppresses ESCC cell proliferation, migration, and invasion; enforces cell cycle arrest and reverses epithelial-mesenchymal transition (EMT). Re-expression of wild-type HSP90AA1 rescues cancer cell growth and EMT in vitro and in vivo, whereas the methylation-deficient R182A mutant does not. Furthermore, clinically oriented analysis shows HSP90AA1 overexpression in ESCC tissues and indicates moderate prognostic utility. Moreover, dual targeting of the PRMT5/HSP90AA1 axis is therapeutically attractive, and combined PRMT5/HSP90AA1 inhibition acts synergistically to exert anti-tumor activity, markedly reducing clonogenicity, invasion, and tumor burden in both cell-derived and patient-derived xenograft models. These findings demonstrate that PRMT5/HSP90AA1 serve as mechanistic drivers and actionable biomarkers in ESCC and provide a rationale for biomarker-guided co-inhibition of PRMT5/HSP90AA1 in translational studies.

cancer biology↗

bioRxiv2025

Disease-associated mutations in the STAT5B SH2 domain reprogram hepatic cholesterol and lipid metabolism

Growth hormone (GH) signaling through STAT5B is a central regulator of hepatic metabolism, yet the functional consequences of disease-associated STAT5B variants remain poorly understood. Here, we analyzed mice carrying STAT5BY665F (gain-of-function) and STAT5BY665H (loss-of-function) variants and dissect their impact on metabolic regulation. STAT5BY665F mice developed hepatic lipid accumulation, hypercholesterolemia, and enhanced insulin sensitivity, whereas STAT5BY665H mice displayed reduced body weight and impaired insulin responsiveness. Transcriptomic analyses revealed that STAT5BY665F activated lipid, cholesterol, and immune transcriptional programs, while STAT5BY665H failed to induce these pathways. Notably, STAT5BY665F substantially feminized male liver gene expression, inducing 77% of female-biased genes while repressing 51% of male-biased genes, thereby mimicking the persistent STAT5B activation characteristic of female livers. ChIP-seq demonstrated extensive STAT5BY665F enhancer occupancy at metabolic and immune loci, contrasting with the minimal chromatin engagement of STAT5BY665H. Beyond the liver, STAT5BY665F broadly reprogrammed adipose tissue gene expression, activating lipid metabolism and immune regulatory networks, whereas STAT5BY665H exerted more restricted effects. Together, these findings illustrate how alterations in STAT5B activity affect enhancer activation and can lead to changes in metabolic function and hepatic sexual dimorphism.

genomics↗

bioRxiv2025

A Chromatin-Structure-Guided Framework for Predictive and Interpretable Regulatory Genomics

Chromatin organization shapes gene regulation by linking distal elements across megabase scales, yet most predictive genomics models still treat the genome as linear, without incorporating three-dimensional structure. Hi-C provides genome-wide chromatin conformation information, but its contact maps are population-averaged, distance-biased, and noisy, obscuring the biologically specific contacts. We present CHROME, a framework built on a self-avoiding polymer ensemble null model that identifies physically specific, non-random Hi-C contacts. By integrating these contacts into graph representations, CHROME enables efficient information transfer across spatially connected loci. It integrates sequence, chromatin accessibility, or pre-trained embeddings into a graph attention architecture to predict cell line-specific ChIP-seq profiles, consistently outperforming local encoder baselines and generalizing to an unseen cell line. The resulting graph embeddings also enhance prediction on tissue-specific eQTL and ClinVar variant pathogenicity, outperforming local sequence-based embeddings. Beyond predictive performance, CHROME provides interpretability through attention-derived neighbor-to-center contributions that reveal how spatially connected loci influence local regulatory activity over multi-megabase distances. Together, these results show that incorporating physically validated chromatin interactions enables more accurate and interpretable modeling of gene regulation and variant effects.

genomics↗

bioRxiv2025

Cold sensing by a glutamate receptor drives avoidance behavior in Drosophila larvae

The ability to sense and avoid noxious environments is essential for animal survival; yet, how this is achieved at the behavioral, neuronal, and molecular levels is not well understood. Here, we use Drosophila larvae as a model to investigate how animals sense and avoid cold temperatures. By implementing custom-built thermoelectric devices capable of delivering rapid and precise thermal stimuli, we find that cold delivered to the larval head evokes robust escape behavioral responses. We identify a group of head-located cold-sensitive neurons as necessary and sufficient for such avoidance responses. We further demonstrate that the kainate-type glutamate receptor Clumsy acts as a novel cold sensor required for head cold sensitivity. Knockdown of Clumsy in head cold-sensing neurons suppresses their cold sensitivity. Heterologous expression of Clumsy confers cold sensitivity. Our results show that Drosophila larvae have evolved the capacity to detect and avoid cold temperatures through a previously uncharacterized cold-sensing mechanism.

neuroscience↗

bioRxiv2025

ChromPolymerDB: A High-Resolution Database of Single-Cell 3D Chromatin Structures for Functional Genomics

The three-dimensional (3D) organization of chromatin plays a critical role in regulating gene expression and genomic processes like DNA replication, repair, and genome stability. Although these processes occur at the individual-cell level, most chromatin structure data are derived from population-averaged assays, such as Hi-C, obscuring the heterogeneity of single-cell conformations. To address this limitation, we developed a polymer physics-based modelling framework, the Sequential Bayesian Inference Framework (sBIF), that deconvolutes bulk Hi-C data to reconstruct single-cell 3D chromatin conformations. To support a broader use of sBIF, we created ChromPolymerDB, a publicly accessible, high-resolution database of single-cell chromatin structures inferred by sBIF. The database contains [~]108 reconstructed 5 kb-resolution single cell structures, spanning over 60,000 genomic loci across 50 human cell types and experimental conditions. ChromPolymerDB features an interactive web interface with tools for 3D structural analysis and multi-omics integration. Users can explore associations between chromatin conformation and gene expression, epigenetic modifications, and regulatory elements. The platform also supports comparative analyses to identify structural changes across cell types, developmental stages, or disease contexts. ChromPolymerDB offers a unique resource for researchers studying the relationship between genome architecture and gene regulation, and for advancing comparative 3D genomics. ChromPolymerDB is available online at https://chrompolymerdb.bme.uic.edu/. GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=85 SRC="FIGDIR/small/678816v1_ufig1.gif" ALT="Figure 1"> View larger version (27K): org.highwire.dtl.DTLVardef@110c807org.highwire.dtl.DTLVardef@bc713forg.highwire.dtl.DTLVardef@1af0ea9org.highwire.dtl.DTLVardef@1eba638_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗

bioRxiv2025

Programming Microbe-Computer Interaction for Customized Life-Nonlife and Cross-Life Communication

This extension of human-computer interaction (HCI) paradigm to the non-intelligent microbe-computer interaction (MCI) is attractive for building generalized life-nonlife communication. Main challenges lie in how to construct biohybrid interface within individual MCI unit and thereafter program cross-life communication based on these MCI units. Here, we propose the design and construction of opto-interfaced MCI unit composed of engineered microbes and computer terminals, for enabling quantifiable transduction between microbial signals and digital signals. These units were able to serve as biohybrid sensors, actuators and transceivers, supporting biomonitoring, bio-actuation and microbe-centric bidirectional operations. Based on these functional units, we established programmable cross-life communication architectures capable of unicast, multicast, and logic-gate-driven communication with temporal dynamic control. The architectures were validated in biomedical scenarios including tumor hypoxia biomarking, epidermal heavy metal detoxification, and colitis intervention, demonstrating robust performance in closed-loop diagnostic-therapeutic applications. Additionally, we integrated MCI with an unmanned surface vessel for environmental monitoring, highlighting its cross-domain potential. Collectively, we provide a scalable strategy for building customized life-nonlife and cross-life communication, which is expected as an important complement to internet of things (IoT).

synthetic biology↗

bioRxiv2025

Neurotype matching in monogamous rodents is modulated by early-life sleep experience

Studies of human sociability indicate stronger social affinity in matched-neurotype dyads (e.g., two individuals with autism or two without) compared to mixed-neurotype dyads (e.g., one individual with autism paired with one without). Is this neurotype matching phenomenon also quantifiable in non-human animals? Using deep learning tools, we analyzed dyadic male-female interactions in prairie voles, a highly social rodent species. To simulate "neurotypes", voles were exposed to either control conditions or early-life sleep disruption (ELSD) during a critical neurodevelopmental period (postnatal days 14-21), recapitulating two features of human autism: developmental sleep disruption and later-life atypical sociability. Analogous to human studies, voles showed signs of reduced social affinity in mixed dyads compared to matched dyads, including sex-specific changes in aggression and body orientation toward the conspecific. These findings advance our understanding of social affinity, providing a framework for new studies in both animal models and humans.

animal behavior and cognition↗

bioRxiv2025

Automated analysis of C. elegans behavior by LabGym: an open-source, AI-powered platform

The genetic tractability, well-mapped circuitry, and diverse behavioral repertoire of the nematode C. elegans make it an ideal model for physiological and behavioral studies. A wide range of methods has been developed for analyzing C. elegans behaviors, evolving with advances in technology such as videography and computer-assisted analysis. Here, we introduce LabGym--an open-source, artificial intelligence (AI)-based platform we recently developed--to the C. elegans research community. We trained deep learning models in LabGym capable of automatically categorizing and quantifying multiple user-defined parameters of worm locomotion behavior in multi-worm videos with high accuracy. Furthermore, we demonstrated their efficacy in quantifying locomotion changes in aging worms. Our work offers a cost-effective, user-accessible approach to behavioral analysis in C. elegans.

animal behavior and cognition↗

bioRxiv2025

Telomere 3'-overhang attrition and persistent telomeric DNA damage response in failing human hearts

BackgroundTelomere homeostasis is critical for normal cellular and organ function, and its dysregulation is implicated in aging and chronic diseases. Although telomere length (TL) is critical for normal telomere function, its functional status can also be altered by many other factors. The organization and protective status of telomere in failing human heart (FHH) remains incompletely understood. MethodsUsing left ventricular tissues from patients with idiopathic dilated cardiomyopathy (IDC), ischemic heart disease (IHD), and non-failing heart (NFH), we performed a comprehensive analysis that included measurements of overall TL and 3 overhang length in left ventricles and cardiomyocytes (CM), assessment of telomere-binding protein associations, and transcriptomic profiling through RNA sequencing. ResultsAlthough TL varied among individuals, reduced median TL was observed only in IDC, while both IDC and IHD showed an increased frequency of very short telomeres, falling below the 5th percentile, in CMs. Strikingly, significant telomere 3-overhang attrition was detected in both disease groups and strongly correlated with elevated H2AX phosphorylated on serine 139 ({gamma}H2AX) in heart tissues, indicating DNA damage. This was accompanied by persistent activation of ataxia telangiectasia mutated (ATM) protein-mediated DNA damage responses and the formation of telomere dysfunction-induced foci (TIFs) in CMs from FHH. Concomitantly, the association of telomeres with the single-stranded telomere-binding protein, Protection of Telomeres 1 (POT1), and the double-stranded telomere-interacting protein, telomere repeat-binding factor 2 (TERF2), was markedly reduced, accompanied by an increase in association of {gamma}H2AX and ssDNA binding protein, RPA with telomeres in IDC and IHD relative to NFH, signifying telomere de-protection. ConclusionsCM telomere dysfunction, characterized by 3 overhang attrition and de-protection, is a common feature in FHH, leading to persistent DNA damage response in telomeres. Better understanding of telomere biology throughout the progression of different heart diseases to heart failure will provide more effective prevention and treatment strategies. GRAPHICAL ABSTRACTHuman telomeres safeguard chromosome ends through a higher-order telomere-loop (t-loop) structure and the shelterin complex, preventing their recognition as DNA double-strand breaks (DSBs). In the failing human heart (FHH) with idiopathic dilated cardiomyopathy (IDC) and ischemic heart disease (IHD), cardiomyocyte telomeres lose protection due to disruption of the shelterin complex, driven by oxidative DNA damage from reactive oxygen species (ROS) targeting G-rich telomeric sequences. This leads to t-loop unfolding, DDR cascade activation, 3 overhang excision, and persistent DNA damage responses with Telomere Dysfunction-Induced Foci (TIFs) formation at the telomeric regions of cardiomyocytes in FHH. The ATM-mediated DDR, along with engagement of both classical non-homologous end joining (cNHEJ) and homologous recombination (HDR) repair pathways, is integral to FHH progression. This study underscores the significance of telomere 3 overhang attrition in the pathogenesis of FHH with IDC and IHD. Unlike changes in overall telomere length, 3 overhang shortening directly correlates with increased DNA damage in human myocardial tissue. Preventing telomere 3 overhang attrition and oxidative damage offers a promising avenue for attenuating heart disease progression toward heart failure. In particular, selective ATM inhibition, as opposed to ATR inhibition, presents a potential therapeutic strategy for patients with IDC and IHD. Future efforts to pinpoint deficiencies in cNHEJ and HDR pathways in failing human hearts may reveal novel targets for heart failure prevention and treatment, providing new strategies in the management of this patient population. CMs, cardiomyocytes; DDR, DNA double-strand break repair; FHH, failing human heart; ROS, reactive oxygen species; t-loop, telomere-loop. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=120 SRC="FIGDIR/small/669993v1_ufig1.gif" ALT="Figure 1"> View larger version (43K): org.highwire.dtl.DTLVardef@1d21937org.highwire.dtl.DTLVardef@757f54org.highwire.dtl.DTLVardef@1fa7472org.highwire.dtl.DTLVardef@1f2e961_HPS_FORMAT_FIGEXP M_FIG C_FIG

molecular biology↗

bioRxiv2025

Metabolomics Ex/GWAS in the Amish reveals novel insights into cardiometabolic disease pathways

We conducted an exome and genome-wide association study (Ex/GWAS) of 1,015 metabolites in serum samples from 5,981 Amish adults. We identified 149 functional or likely functional genetic variants, based on CADD scores significantly associated (P < 5.8 x 10-8) with 519 metabolite levels, 69 of which were >10-fold enriched in the Amish versus Europeans in gnomAD. We discovered novel associations involving a PCK2 splice-donor variant (rs138881435) and metabolites important for energy metabolism and mitochondrial function. In UK Biobank participants, this variant was associated with increased longitudinal relaxation time (T1) values on liver MRI suggesting increased inflammation and fibrosis. Furthermore, we found novel associations involving ENPEP stop-gain variant (rs33966350) and peptide and amino acid metabolite levels, implicating ENPEPs roles in blood pressure regulation. Additional known and novel genetic variant-metabolite associations were identified including genes involved in Amish-enriched Mendelian diseases and cardiometabolic traits, providing insights into underlying mechanisms of these disorders.

genomics↗

bioRxiv2025

An improved FLARE system for recording and manipulating neuronal activity

Recording and manipulating neuronal ensembles that underlie cognition and behavior in vivo is challenging. FLARE is a light- and calcium-gated transcriptional reporting system for labeling activated neurons on the order of minutes. However, FLARE is limited by its sensitivity to prolonged neuronal activities. Here, we present an improved version of FLARE, termed cytoFLARE. cytoFLARE incorporates cytosolic expression of the transcription factor and a more sensitive pair of calcium sensing domains. We showed that cytoFLARE provides more calcium- and light-dependent signals in HEK293T cells and higher signal-to-background ratios in neuronal cultures. We further established cytoFLARE transgenic Drosophila models and applied cytoFLARE to label activated neurons upon sensory or optogenetic stimulation within a defined time window. Notably, through cytoFLARE-driven expression of an optogenetic actuator, we successfully reactivated neurons involved in the larval nociceptive system. Our findings demonstrate the first characterization and application of time-gated calcium integrators in Drosophila.

bioengineering↗

bioRxiv2024

Bridging the Gap in Immunotherapy Prediction: The AGAE Score as a Pan Cancer Biomarker for Immune Checkpoint Inhibitor Response

BackgroundImmune checkpoint inhibitor (ICI) therapy efficacy varies among cancer patients, necessitating precise predictive biomarkers for optimized treatment strategies. MethodsWe developed the Adaptive best subset selection algorithm and Genetic algorithm Aided Ensemble learning (AGAE) score through multi-cohort transcriptomic analysis of ICI-treated patients. The AGAE score incorporated gene-pairing, Adaptive Best Subset Selection for feature optimization, and a Genetic Algorithm for optimal basic learner identification. We explored correlations between AGAE score and immune microenvironments using multi-omics data. Potential targets were screened using 17 CRISPR datasets and validated through in vitro and in vivo experiments. ResultsThe AGAE score demonstrated robust predictive power for ICI therapy outcomes, with lower scores correlating with enhanced treatment response. The AGAE score outperformed published signatures and conventional biomarkers. Lower AGAE scores were associated with increased immune cell infiltration, higher immunogenicity, and enhanced antitumor immune activity. The CEP55 was identified as a potential key target driving immune evasion through AGAE scoring and CRISPR screening. Experimental validation showed CEP55 downregulation attenuated tumor cell malignancy and augmented ICI therapy efficacy by modulating T cell responses. ConclusionsThe AGAE score was a potent predictor of ICI therapy efficacy, facilitating refined patient stratification. CEP55s role in the tumor microenvironments immune response highlights its potential as a therapeutic target. Targeted interventions against CEP55 may improve immunotherapy precision.

cancer biology↗

bioRxiv2024

Interpretable high-resolution dimension reduction of spatial transcriptomics data by DeepFuseNMF

Spatial transcriptomics (ST) technologies have revolutionized tissue architecture studies by capturing gene expression with spatial context. However, high-dimensional ST data often have limited spatial resolution and exhibit considerable noise and sparsity, posing significant challenges in deciphering subtle spatial structures and underlying biological activities. Here, we introduce SpaHDmap, an interpretable dimension reduction framework that enhances spatial resolution by integrating ST gene expression with high-resolution histology images. SpaHDmap incorporates non-negative matrix factorization into a multimodal fusion encoder-decoder architecture, enabling the identification of interpretable, high-resolution embeddings. Furthermore, SpaHDmap can simultaneously analyze multiple samples and is compatible with various types of histology images. Extensive evaluations on synthetic and real ST datasets from various technologies and tissue types demonstrate that SpaHDmap can effectively produce highly interpretable, high-resolution embeddings, and detects refined spatial structures. SpaHDmap represents a powerful approach for integrating ST data and histology images, offering deeper insights into complex tissue structures and functions.

bioinformatics↗

bioRxiv2024

Quantifying social roles in multi-animal videos using subject-aware deep-learning

Analyzing social behaviors is critical for many fields, including neuroscience, psychology, and ecology. While computational tools have been developed to analyze videos containing animals engaging in limited social interactions under specific experimental conditions, automated identification of the social roles of freely moving individuals in a multi-animal group remains unresolved. Here we describe a deep-learning-based system - named LabGym2 - for identifying and quantifying social roles in multi-animal groups. This system uses a subject-aware approach: it evaluates the behavioral state of every individual in a group of two or more animals while factoring in its social and environmental surroundings. We demonstrate the performance of subject-aware deep-learning in different species and assays, from partner preference in freely-moving insects to primate social interactions in the field. Our subject-aware deep learning approach provides a controllable, interpretable, and efficient framework to enable new experimental paradigms and systematic evaluation of interactive behavior in individuals identified within a group.

animal behavior and cognition↗

bioRxiv2024

Cognitive behavioral phenotyping of DSCAM heterozygosity as a model for autism spectrum disorder

It is estimated that 1 in 36 children are affected by autism spectrum disorder (ASD) in the United States, which is nearly a twofold increase from a decade ago. Recent genetic studies have identified de novo loss-of-function (dnLoF) mutations in the Down Syndrome Cell Adhesion Molecule (DSCAM) as a strong risk factor for ASD. Previous research has shown that DSCAM ablation confers social interaction deficits and perseverative behaviors in mouse models. However, it remains unknown to what extent DSCAM underexpression captures the full range of behaviors, specifically cognitive phenotypes, presented in ASD. Here, we conducted a comprehensive cognitive behavioral phenotyping which revealed that loss of one copy of DSCAM, as in the DSCAM2J+/- mice, displayed hyperactivity, increased anxiety, and motor coordination impairments. Additionally, hippocampal-dependent learning and memory was affected, including working memory, long-term memory, and contextual fear learning. Interestingly, implicit learning processes remained intact. Therefore, DSCAM LoF produces autistic-like behaviors that are similar to human cases of ASD. These findings further support a role for DSCAM dnLoF mutations in ASD and suggest DSCAM2J+/- as a suitable model for ASD research. Summary StatementAutism spectrum disorder represents a growing patient population. Loss of one copy of the DSCAM gene provides a promising mouse model that reproduces autistic-like behaviors for research and therapeutic testing.

neuroscience↗

bioRxiv2024

CASTpFold: Computed Atlas of Surface Topography of the universe of protein Folds

Geometric and topological properties of protein structures, including surface pockets, interior cavities, and cross channels, are of fundamental importance for proteins to carry out their functions. Computed Atlas of Surface Topography of proteins (CASTp) is a widely used web server for locating, delineating, and measuring these geometric and topological properties of protein structures. Recent developments in AI-based protein structure prediction such as AlphaFold2 (AF2) have significantly expanded our knowledge on protein structures. Here we present CASTpFold, a continuation of CASTp that provides accurate and comprehensive identifications and quantifications of protein topography. It now provides (i) results on an expanded database of proteins, including the Protein Data Bank (PDB) and non-singleton representative structures of AlphaFold2 structures, covering 183 million AF2 structures; (ii) functional pockets prediction with corresponding Gene Ontology (GO) terms or Enzyme Commission (EC) numbers for AF2-predicted structures; and (iii) pocket similarity search function for surface and protein-protein interface pockets. The CASTpFold web server is freely accessible at https://cfold.bme.uic.edu/castpfold/.

biophysics↗