Biology subjects

Zheng, M.

Publications and source records attributed to Zheng, M..

At least 37 records · Page 2Linked to original sources

bioRxiv2025

Benchmarking Real-World Applicability of Molecular Generative Models from De novo Design to Lead Optimization with MolGenBench

Structure-based drug design (SBDD) has been profoundly reshaped by the advent of deep generative models, yet their practical impact on drug discovery remains limited. A central issue is the absence of a rigorous, application-oriented benchmark that mirrors the multi-stage, target-aware workflows of real-world pharmaceutical development. Inspired by recent advances in benchmarking for computer vision and large language models, where systematic evaluation has catalysed rapid progress, we introduce MolGenBench, a comprehensive benchmark designed to close the gap between molecular generation algorithms and tangible drug discovery outcomes. MolGenBench integrates a structurally diverse, large-scale dataset spanning 120 protein targets, 5,433 chemical series comprising 220,005 experimentally confirmed active molecules. Beyond conventional de novo generation, it incorporates a dedicated hit-to-lead (H2L) optimization scenario, which represents a critical phase in hit optimization that is seldom evaluated in existing benchmarks. We further introduce novel, pharmaceutically grounded metrics that assess a models ability to both rediscover target-specific actives or progressively optimize compounds for potency. Through extensive evaluation, MolGenBench uncovers significant gaps between current generative models and the demands of real-world drug development, establishing a foundational resource for building generative models with enhanced practical impact and accelerated translational potential.

bioinformatics↗

bioRxiv2025

Broad-Spectrum HIV-1 Detection and Neutralization via Multivalent Designer DNA Nanostructures

Early and accurate detection of HIV-1 remains a critical unmet need, particularly during the acute phase of infection when viral loads are low and transmission risk is highest. Here, we report a modular diagnostic and antiviral platform based on designer DNA nanostructures engineered for high-affinity recognition of the HIV-1 envelope glycoprotein (GP120). A custom DNA aptamer, termed HINT, was developed to bind GP120 across major HIV-1 subtypes (Groups M and P; subtypes A and B) with nanomolar affinity. To amplify binding strength, HINT aptamers were spatially patterned onto a net-shaped DNA nanostructure (DNA-NetHINT) that geometrically matches the trimeric GP120 spikes on the viral surface. Using multivalent interactions, the nanostructure enabled up to 104-fold improvement in binding affinity (sub-picomolar KD), confirmed by surface plasmon resonance. Integration of DNA-NetHINT into a paper-based lateral flow assay produced a low-cost, saliva-compatible self-testing device capable of detecting intact HIV-1 virions at concentrations as low as 328 viral copies per test, outperforming commercial fourth-generation rapid diagnostic tests. In addition to its diagnostic capabilities, the DNA-NetHINT construct exhibited potent antiviral activity, reducing pseudovirus infection with an EC50 of [~]1.8 nM, nearly 1,000-fold more effective than free aptamers. This work demonstrates a dual-function DNA nanotechnology platform that enables both ultrasensitive HIV-1 detection and entry inhibition. The approach is broadly applicable to other enveloped viruses and represents a promising step toward next-generation molecular theranostics for infectious disease management.

bioengineering↗

bioRxiv2025

Endothelin Signaling via EDNRB receptor Reduces Proliferation and Promotes Proneural-to-Mesenchymal Transition in Gliomas

Diffuse gliomas are incurable primary brain tumors encompassing three histo-molecular subtypes: glioblastomas (GB), astrocytomas, and oligodendrogliomas. The latter two harbor IDH1 mutations and exhibit slower progression than glioblastomas. Diffuse gliomas are composed of highly plastic tumor cells capable of transitioning between astrocyte-like, oligodendrocyte-like, progenitor-like, and mesenchymal-like states, driven by genetic alterations and microenvironmental cues. The proneural-to-mesenchymal transition (PMT), associated with increased malignancy, is notably influenced by cytokines in the tumor microenvironment. Endothelin cytokines (ET-1, ET-2, ET-3), primarily secreted by vascular cells, regulate not only vascular tone but also astrocyte and neural stem cell proliferation via the G-protein-coupled receptors EDNRA and EDNRB. Prior studies using serum-cultured glioma lines suggested pro-proliferative effects of endothelins; however, such models poorly recapitulate the in vivo glioma context. In this study, we comprehensively revisited endothelin signaling - covering receptor expression, regulation, downstream pathways, and cellular responses-using eleven serum-free, patient-derived glioma lines (glioblastomas, IDH-wt and IDH-mutant oligodendrogliomas and astrocytomas), along with primary tumor samples. Multi-omics and electrophysiological analyses revealed EDNRB as the predominant receptor, enriched in astrocyte-like cells, upregulated by BMPs or growth factor withdrawal, and downregulated by interferons, IL-6 cytokines, endothelins, and Hippo/YAP activation. In contrast, EDNRA was expressed by a perivascular tumor subpopulation and induced by Notch signaling in glioblastomas but not in IDH1-mutant cells. Functionally, endothelins reduced proliferation across all models while promoting migration and PMT. Mechanistically, EDNRB activation increased intracellular Ca{superscript 2} and activated ERK, STAT3, and apamin-sensitive SK2/SK3 potassium channels. These findings identify endothelin signaling as an important regulator of glioma cell plasticity and behavior. HighlightsO_LIEDNRB is the predominant endothelin receptor expressed in glioma cells, with a small subset of tumor cells expressing EDNRA in close proximity to blood vessels C_LIO_LIEndothelin signaling reduces proliferation while promoting cell migration and Proneural-to-Mesenchymal transition C_LIO_LIEndothelin activates downstream Ca2+, K+, ERK, and STAT3 signaling pathways C_LIO_LIEDNRB expression is both positively and negatively regulated by inflammatory cytokines and the Hippo/YAP1 pathway, whereas EDNRA is upregulated by Notch signaling and hypoxia C_LI

cancer biology↗

bioRxiv2025

Advancing Ligand Binding Affinity Prediction with Cartesian Tensor-Based Deep Learning

0.We present PBCNet2.0, a cartesian tensor-based Siamese Neural Network for protein-ligand relative binding affinity prediction. Trained on 8.6 million protein-ligand complex structure pairs, PBCNet2.0 achieves zero-shot performance comparable to computationally intensive physics-based simulations. Our prioritization experiments show that PBCNet2.0 speeds up binding affinity optimization by 718% while reducing resource use by 41%. Through extensive retrospective experiments, we demonstrate that PBCNet2.0 intrinsically comprehends protein-ligand interactions, showing high sensitivity to intermolecular interactions and exceptional perception of spatial geometric information. Strikingly, PBCNet2.0 exhibits an emergent capability to predict affinity changes induced by binding residue variations, highlighting its potential for identifying resistance mutation. We prospectively validated these capabilities on two targets ENPP1 and ALDH1B1, where PBCNet2.0 successfully identified affinity shifts arising from subtle molecular interactions and conformational differences, and pinpointed critical binding residues with an 83% hit rate. This combination of computational efficiency, spatial geometric perception of binding site, and generalizable affinity prediction establishes PBCNet2.0 as a transformative tool for developing pharmacological probes for all human proteins.

molecular biology↗

bioRxiv2025

Central nervous system atlas of larval zebrafish constructed using the morphology of single excitatory and inhibitory neurons

Single-neuron morphology mapping is fundamental for deciphering brain architecture and function, yet in vertebrates it remains limited in spatial coverage and cell-type information. Here we present a nervous system-wide single-neuron morphology atlas of larval zebrafish with annotated excitatory/inhibitory identity and dendrite-axon polarity. The atlas comprises >20,000 reconstructed neurons from >13,400 animals, spanning the central and peripheral nervous system and representing [~]24.5% of the total excitatory and inhibitory neurons in the brain, all integrated into a common physical space with neuroanatomical parcellations. We define >500 excitatory and inhibitory neuronal morphotypes and construct putative inter-cell-type and interregional connectomes. Projection and morphotype analyses reveal distinct laterality and directionality biases and modular innervation patterns in excitatory and inhibitory neurons, while network analyses uncover fundamental connectivity motifs, hubs, and sensorimotor pathways. Accessible through custom-built interactive platform ZExplorer, this atlas provides a foundational resource for multimodal data integration, circuit analysis, computational modeling, and hypothesis-driven research.

neuroscience↗

bioRxiv2025

Controllable Gap Junctions by Vitamin B12 and Light

Gap junctions, mediate rapid signal transduction between contiguous cells, which are indispensable for multicellular organisms to coordinate cellular activities across numerous physiological processes. However, precise control of gap junctions remains elusive. Herein, we present CarGAP, a single-component chemo-optogenetic tool that utilizes the C-terminal adenosylcobalamin (AdoB12) binding domain of a photoreceptor protein (i.e., CarHC) to achieve reversible control over both vertebrate and invertebrate gap junctions with spatiotemporal precision. The vertebrate CarGAP (i.e., Cx-CarGAP), created by genetically fusing connexins with CarHC in mammalian cells, can efficiently block the gap junction channels through AdoB12-induced protein oligomerization, and subsequently reinstate them via green light-induced protein disassembly. We further introduced the CarGAP system (i.e., Inx-CarGAP) to the Drosophila ovary, enabling reversible control over the heterotypic gap junctions formed by innexin2 (Inx2) and innexin4 (Inx4, also known as zero population growth, Zpg), thereby uncovering the roles of gap junctions in stem cell-niche interactions. This study illustrates CarGAP as a generalizable chemo-optogenetic tool for interrogating the functions of gap junctions in various biological contexts.

bioengineering↗

bioRxiv2025

Heterobifunctional Protein Binders Enable Cell Type-Specific Killing Through In-cell Enrichment

Non-catalytic heterobifunctional molecules promise to expand the range of therapeutic options by establishing complexes between key target proteins and accessory presenter proteins equipped with additional properties. Here, we systematically investigate the rational design of such molecules, explore the biochemical basis of complex formation and determine how they achieve cellular efficacy using the endogenously expressed immunophilin FKBP12 and the transcriptional regulator BRD4 as paradigms. We present classes of bifunctional molecules that enable selective, FKBP12-dependent killing of specific cell types at subnanomolar concentrations and allow to differentiate between closely similar bromodomains. We propose that the strongly potentiated efficacy of these bifunctional compounds is based on cellular enrichment through binding to the highly abundant presenter protein FKBP12, a mechanism we term "CellTrap". Our findings substantiate the concept that highly expressed, non-essential proteins can be repurposed as selective recruiters to expand therapeutic windows of existing small-molecule inhibitors, opening new avenues for designing targeted drugs with improved cell-type specificity.

biochemistry↗

bioRxiv2025

An approach to differentiating biological from chemical sulfide oxidation on corroded concrete sewer surfaces

Hydrogen sulfide (H2S)-induced sewer corrosion largely impacts the longevity of infrastructure worldwide. The corrosion incorporates two pathways, i.e., chemical and biological oxidation of H2S to sulfuric acid; however, their rates are difficult to differentiate on corroded sewer pipelines, limiting understanding and application of effective strategies for corrosion control. Here, this study develops and reports a kinetic approach to determining rates of chemical and biological sulfide oxidation. By measuring the total sulfide uptake rate of sewer corrosion concretes, extensive experimental data was obtained for calibration and validation of the proposed kinetic model. The modelling results fit with measured data indicating that both chemical and biological sulfide oxidation can be well-described by the developed model. Model predictions revealed that along with sewer corrosion (i.e., pH decrease), chemical sulfide oxidation rate was significantly accelerated, along with an increase in biological rate. This result suggests that the chemical sulfide oxidation process should not be overlooked for seriously corroded concretes, and synergistic strategies to suppress both biological and chemical rates are required. Specifically, the chemical process can predominate sulfide oxidation with a gas-H2S concentration higher than 300 ppm. The work provides a useful tool for future corrosion studies and also valuable insights for sewer corrosion management practices. SynopsisThis study reports a kinetic approach to determining chemical and biological H2S oxidation rates of sewer concretes. Graphic abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=125 SRC="FIGDIR/small/653906v2_ufig1.gif" ALT="Figure 1"> View larger version (31K): org.highwire.dtl.DTLVardef@1cfd16aorg.highwire.dtl.DTLVardef@bd9677org.highwire.dtl.DTLVardef@e1610forg.highwire.dtl.DTLVardef@ab3dfc_HPS_FORMAT_FIGEXP M_FIG C_FIG

microbiology↗

bioRxiv2025

Bridging 3D Molecular Structures and Artificial Intelligence by a Conformation Description Language

Artificial intelligence, particularly language models (LMs), is reshaping research paradigms across scientific domains. In the fields of chemistry and pharmacy, chemical language models (CLMs) have achieved remarkable success in two-dimensional (2D) molecular modeling tasks by leveraging one-dimensional (1D) representations of molecules, such as SMILES and SELFIES. However, extending these successes to three-dimensional (3D) molecular modeling remains a significant challenge, largely due to the absence of effective 1D representations for capturing 3D molecular structures. To address this gap, we introduce ConfSeq, a novel molecular conformation description language that integrates SMILES with internal coordinates including dihedral angles, bond angles, and pseudo-chirality. This design naturally ensures SE(3) invariance, while preserving the human readability and conciseness characteristic of SMILES. ConfSeq enables the reformulation of a range of 3D molecular modeling tasks, such as molecular conformation prediction, 3D molecular generation, and 3D molecular representation, into sequence modeling problems. Then, by simply employing a standard Transformer architecture, we achieve state-of-the-art performance on various benchmark sets. Furthermore, compared to widely used diffusion-based approaches in 3D molecular modeling, the ConfSeq-based method offers unique advantages in inference efficiency, generation controllability, and enables scoring of generated molecules. We believe that ConfSeq can serve as a foundational tool, advancing the development of sequence-based 3D molecular modeling methods.

bioinformatics↗

bioRxiv2025

EasyMultiProfiler: An Efficient Multi-Omics Data Integration and Analysis Workflow for Microbiome Research

O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=141 SRC="FIGDIR/small/649266v1_ufig1.gif" ALT="Figure 1"> View larger version (60K): org.highwire.dtl.DTLVardef@14dd0b1org.highwire.dtl.DTLVardef@1c04a79org.highwire.dtl.DTLVardef@159894dorg.highwire.dtl.DTLVardef@1454a01_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOGraphical AbstractC_FLOATNO C_FIG HighlightsO_LIEasyMultiProfiler provides unified integration of microbial multi-omics data C_LIO_LIEasyMultiProfiler enables natural language-style analytical workflows C_LIO_LIEasyMultiProfiler solves key challenges in standardization and reproducibility C_LIO_LIEasyMultiProfiler extracts meaningful biological insights efficiently C_LI Host-microbiome interactions are crucial in maintaining physiological homeostasis and influencing disease progression. While traditional microbiome research, focused on microbial diversity and abundance, provided essential frameworks, the integration of multi-omics approaches established a far more comprehensive, systems-level understanding of microbiome functionality. However, significant methodological challenges in multi-omics data integration remain, including inconsistent sample coverage, heterogeneous data formats, and complex downstream analytical workflows. These challenges impact the reproducibility and reliability of results and underscore the critical requirement for developing standardized, systematic workflows. Thus, we developed the EasyMultiProfiler (EMP) workflow, a streamlined and efficient analytical framework for multi-omics data analysis. EMP provides a comprehensive infrastructure based on the SummarizedExperiment and MultiAssayExperiment classes, creating a unified framework for storing and analyzing omics data. The frameworks architecture comprises five interconnected functional modules: data extraction, preparation, support, analysis, and visualization. These modules are smoothly integrated into natural language-style analytical workflows, offering users an efficient and standardized solution for multi-omics analysis. EMP addresses critical challenges in multi-omics data analysis, including data integration, workflow standardization, and result reproducibility. The platforms modular architecture and intuitive interface provide researchers and clinicians with a robust, flexible workflow to systematically extract biologically relevant insights from complex multi-omics datasets.

bioinformatics↗

bioRxiv2025

Two telomere-to-telomere, gap-free genome assemblies and comparisons revealed the conserved key genes associated with sugar accumulation in Rubus genus

For the first time, we assembled two highly continuous, completely gap-free reference genomes of the Rubus subgenus: Rubus hirsutus Thunb. Penglei (XMM) and Rubus eustephanos Focke ex Diels Dahongpao (DHP), which are widely distributed in southern China with similar phenotypic traits (Figures 1A, 1B, and Figure S1), yet ripe fruits display distinct sugar accumulation levels (Table S1), making them ideal candidates for investigating the mechanisms underlying sugar accumulation in the Rubus genus. The XMM (213.53 Mb, 28,204 genes) and DHP (218.26 Mb, 28,569 genes) genomes exhibit close evolutionary relationships, diverging approximately 3.21 Mya. Comparative genomics identified extensive synteny, interspecific structural variations (translocations, inversions, segmental duplications), and presence/absence variation (PAV). Using Hi-C interaction heatmaps and Sanger sequencing, we validated interspecific structural inversions. Additionally, we identified a sugar transporter gene (MFS1), which is present in XMM but absent in DHP. Combined analysis of the gene family expansion/contraction and transcriptome identified two conserved key genes (RhSTP13 and RhSTP7) associated with sugar accumulation in Rubus genus and displayed distinct roles through transient expression assay. To facilitate functional genomics study, we also established a comprehensive Rubus database, RubusDB, a freely accessible repository consolidating all genomic, transcriptomic and phenotypic data of Rubus genus. These findings provide a foundational framework for elucidating the genetic basis of sugar accumulation, genome diversification, and trait improvement in Rubus species. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=170 SRC="FIGDIR/small/646780v2_fig1.gif" ALT="Figure 1"> View larger version (72K): org.highwire.dtl.DTLVardef@1baca20org.highwire.dtl.DTLVardef@2c01eaorg.highwire.dtl.DTLVardef@131dc00org.highwire.dtl.DTLVardef@62a04f_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOFIGURE 1.C_FLOATNO Phenotypic characterization, de novo genome assemblies, variation, and evolution of Rubus hirsutus Thunb. Penglei (XMM) and Rubus eustephanos Focke ex Diels Dahongpao (DHP). (A) Illustration of XMM whole plants, stem, flower bud and flower, respectively. (B) Illustration of DHP whole plants, stem, flower bud and flower, respectively. (C) The chromosome karyotype analysis of XMM. Bar = 10 {micro}m. (D) Circular diagram of XMM and DHP reference genomes. (a) Chromosomes are represented by centromeres (dark red) and telomeres (blue); (b) CEN18 density; (c) CEN17 density; (d) Gene density; (e) TE density; (f) PAV density; (g) SNP density; (h) Collinear lines at the center of diagram highlight homoeologous chromosomes relationships and non-homoeologous regions. (E) Genomic alignments between XMM and DHP. Inversions, duplications and translocations are marked with orange, blue and green ribbons, respectively. (F) Identification of large inversions in chromosome 7 between XMM and DHP. The three heatmaps show the chromatin interaction matrix, including mapping Hi-C data of DHP against DHP genome (left), mapping Hi-C data of DHP against XMM genome (middle) and mapping Hi-C data of XMM against XMM genome (right). The lower panel illustrates gene alignments between XMM and DHP. (G) Synteny, structural rearrangements, monomer, and identity distribution in the 16-20 Mbp region of chromosome 6 between XMM and DHP. (H) Local genome synteny of chromosome 5, with a structural variation that is present only in XMM. The PAV region includes a gene related to sugar transporter (MFS1, XMM05G025540). (I) Gene-level matchings of MFS genes in the PAVs region between XMM and DHP. (J) Phylogenetic relationships and divergence times between raspberries and other Rosaceae species. The black numbers close to the divergence nodes indicate the divergence times and the red bars represent the 95% confidence intervals. The red and black numbers indicate expanded and contracted ortholog groups at the corresponding node. Scale bar corresponds to 10 Mya. C_FIG

genomics↗

bioRxiv2024

PERK Signaling Maintains Hematopoietic Stem Cell Pool Integrity under Endoplasmic Reticulum Stress by Promoting Proliferation

The integrity of the hematopoietic stem cell (HSC) pool relies on efficient long-term self-renewal and the timely removal of damaged or differentiation-prone HSCs. Previous studies have demonstrated the PERK branch of the unfolded protein response (UPR) drives specific programmed cell death programs to maintain HSC pool integrity in response to ER stress. However, the role of PERK in regulating HSC fate in vivo remains unclear. Here, we demonstrate that PERK is dispensable for normal hematopoiesis and HSC self-renewal under steady-state conditions. In contrast, PERK is activated to promote HSC proliferation and depletion in response to ER stress induced by the inactivation of ER-associated degradation (ERAD), via the knockout of key components of ERAD Sel1L or Hrd1. Inhibition of PERK, either through genetic knockout or knock-in of a point mutation that eliminates PERK kinase activity, significantly restores the HSC defects induced by Sel1L or Hrd1 knockout. Mechanistic studies reveal that ERAD deficiency does not lead to HSC death or ROS accumulation. Instead, PERK promotes the activation of mTOR signaling and drives abnormal proliferation of HSCs, impairing their self-renewal potential. This process removes stressed HSCs, thereby maintaining HSC pool integrity. Our study uncovers a PERK-centered strategy employed by HSCs to preserve their pool integrity independently of apoptosis. Key pointsO_LIPERK is not required for steady-state hematopoiesis but preserves hematopoietic stem cell pool integrity in response to increased ER stress. C_LIO_LIUnder ER stress induced by ERAD deficiency, PERK is activated to promote mTOR signaling and HSC hyper-proliferation, depleting damaged HSCs. C_LI

cell biology↗

bioRxiv2024

MYH11 rare variant augments aortic growth and induces cardiac hypertrophy and heart failure with pressure overload

Smooth muscle cell-specific myosin heavy chain, encoded by MYH11, is selectively expressed in smooth muscle cells (SMCs). Pathogenic variants in MYH11 predispose to a number of disorders, including heritable thoracic aortic disease associated with patent ductus arteriosus, visceral myopathy, and megacystis-microcolon-intestinal hypoperistalsis syndrome. Rare variants of uncertain significance occur throughout the gene, including MYH11 p.Glu1892Asp, and we sought to determine if this variant causes thoracic aortic disease in mice. Genomic editing was used to generate Myh11E1892D/E1892D mice. Wild-type (WT) and mutant mice underwent cardiovascular phenotyping and with transverse aortic constriction (TAC). Myh11E1892D/E1892D and WT mice displayed similar growth, blood pressure, root and ascending aortic diameters, and cardiac function up to 13 months of age, along with similar contraction and relaxation on myographic testing. TAC induced hypertension similarly in Myh11E1892D/E1892D and WT mice, but mutant mice showed augmented ascending aortic enlargement and increased elastic fragmentation on histology. Unexpectedly, male Myh11E1892D/E1892D mice two weeks post-TAC had decreased ejection fraction, stroke volume, fractional shortening, and cardiac output compared to similarly treated male WT mice. Importantly, left ventricular mass increased significantly due to primarily posterior wall thickening, and cardiac histology confirmed cardiomyocyte hypertrophy and increased collagen deposition in the myocardium and surrounding arteries. These results further highlight the clinical heterogeneity associated with MYH11 rare variants. Given that MYH11 is selectively expressed in SMCs, these results implicate a role of vascular SMCs in the heart contributing to cardiac hypertrophy and failure with pressure overload. Author SummaryIn this study, we explore the impact of a specific genetic variant, MYH11 p.Glu1892Asp, on the heart and blood vessels in mice. The MYH11 gene is crucial for smooth muscle cells, which are found in the walls of blood vessels and play an important role in various vascular diseases. We created mice with this genetic variant to see if it would lead to thoracic aortic disease, a condition affecting the main artery from the heart. We found that mice with the variant were similar to normal mice in many aspects, such as growth, blood pressure, and heart function, for up to 13 months. However, when we induced high blood pressure in the mice, the mutant mice showed more significant enlargement of the aorta and damage to the elastic fibers in the aortic walls. Interestingly, male mutant mice also developed heart problems, such as reduced heart pumping ability and increased heart muscle thickness, after the high blood pressure challenge. This was accompanied by signs of heart muscle cell enlargement and increased tissue stiffness. These findings suggest that this rare MYH11 variant can contribute to a range of heart and vascular issues, particularly under conditions of pressure overload, and highlight the importance of smooth muscle cells in the development of these problems.

genetics↗

bioRxiv2024

Dysregulated Airway Host Defense in Hyper IgE Syndrome due to STAT3 Mutations

RationaleHyper IgE syndrome (STAT3-HIES), also known as Jobs syndrome, is a rare immunodeficiency disease typically caused by dominant-negative STAT3 mutations. STAT3-HIES syndrome is characterized by chronic pulmonary infection and inflammation, suggesting impairment of pulmonary innate host defense. ObjectivesTo identify airway epithelial host defense defects consequent to STAT3 mutations that, in addition to reported mutant STAT3 immunologic abnormalities, produce pulmonary infection. MethodsSTAT3-HIES sputum was evaluated for biochemical/biophysical properties. STAT3-HIES excised lungs were harvested for histology; bronchial brush samples were collected for RNA sequencing and in vitro culture. A STAT3-HIES-specific mutation (R382W), expressed by lentiviruses, and a STAT3 knockout, generated by CRISPR/Cas9, were maintained in normal human bronchial epithelia under basal or inflammatory (IL1{beta}) conditions. Effects of STAT3 deficiency on transcriptomics, and epithelial ion channel, secretory, antimicrobial, and ciliary functions were assessed. Measurements and Main ResultsMucus concentrations and viscoelasticity were increased in STAT3-HIES sputum. STAT3-HIES excised lungs exhibited mucus obstruction and elevated IL1{beta} expression. STAT3 deficiency impaired CFTR-dependent fluid and mucin secretion, inhibited expression of antimicrobial peptides, cytokines, and chemokines, and acidified airway surface liquid at baseline and post-IL1{beta} exposure in vitro. Notably, mutant STAT3 suppressed IL1R1 expression. STAT3 mutations also inhibited ciliogenesis in vivo and impaired mucociliary transport in vitro, a process mediated via HES6 suppression. Administration of a {gamma}-secretase inhibitor increased HES6 expression and improved ciliogenesis in STAT3 R382W mutant cells. ConclusionsSTAT3 dysfunction leads to multi-component defects in airway epithelial innate defense, which, in conjunction with STAT3-HIES immune deficiency, contributes to chronic pulmonary infection.

cell biology↗

bioRxiv2024

SciMind: A Multimodal Mixture-of-Experts Model for Advancing Pharmaceutical Sciences

Large language models (LLMs) have made substantial strides, but their use in reliably tackling issues within specialized domains, particularly in interdisciplinary areas like pharmaceutical sciences, is hindered by data heterogeneity, knowledge complexity, unique objectives, and a spectrum of constraint conditions. In this area, diverse modalities such as nucleic acids, proteins, molecular structures, and natural language are often involved. We designed a specialized token set and introduced a new Mixture-of-Experts (MoEs) pre-training and fine-tuning strategy to unify these modalities in one model. With this strategy, weve created a multi-modal mixture-of-experts foundational model for pharmaceutical sciences, named SciMind. This model has undergone extensive pre-training on publicly accessible datasets including nucleic acid sequences, protein sequences, molecular structure strings, and biomedical texts, and delivers good performance on biomedical text comprehension, promoter prediction, protein function prediction, molecular description, and molecular generation.

bioinformatics↗

bioRxiv2024

FAPM: Functional Annotation of Proteins using Multi-Modal Models Beyond Structural Modeling

Assigning accurate property labels to proteins, like functional terms and catalytic activity, is challenging, especially for proteins without homologs and "tail labels" with few known examples. Unlike previous methods that mainly focused on protein sequence features, we use a pretrained large natural language model to understand the semantic meaning of protein labels. Specifically, we introduce FAPM, a contrastive multi-modal model that links natural language with protein sequence language. This model combines a pretrained protein sequence model with a pretrained large language model to generate labels, such as Gene Ontology (GO) functional terms and catalytic activity predictions, in natural language. Our results show that FAPM excels in understanding protein properties, outperforming models based solely on protein sequences or structures. It achieves state-of-the-art performance on public benchmarks and in-house experimentally annotated phage proteins, which often have few known homologs. Additionally, FAPMs flexibility allows it to incorporate extra text prompts, like taxonomy information, enhancing both its predictive performance and explainability. This novel approach offers a promising alternative to current methods that rely on multiple sequence alignment for protein annotation. The online demo is at: https://huggingface.co/spaces/wenkai/FAPM_demo.

bioinformatics↗

bioRxiv2024

A Transcriptomic Hourglass In Brown Algae

Complex multicellularity has emerged independently across a few eukaryotic lineages and is often associated with the rise of elaborate, tightly coordinated developmental processes. How multicellularity and development are interconnected in evolution is a major question in biology. The hourglass model of embryonic evolution depicts how developmental processes are conserved during evolution, predicting morphological and molecular divergence in early and late embryo stages, bridged by a conserved mid-embryonic (phylotypic) period linked to the formation of the basic body plan. Initially found in animal embryos, molecular hourglass patterns have recently been proposed for land plants and fungi. However, whether the hourglass pattern is an intrinsic feature of all developmentally complex eukaryotic lineages remains elusive. Here, we tested the prevalence of a (molecular) hourglass in the brown algae, the third most developmentally complex lineage on earth that has evolved multicellularity independently from animals, fungi, and plants. By exploring the evolutionary transcriptome of brown algae with distinct morphological complexities, we uncovered an hourglass pattern during embryogenesis in developmentally complex species. Filamentous algae without a canonical embryogenesis display an evolutionary transcriptome that is most conserved in multicellular stages of the life cycle, whereas unicellular stages are more rapidly evolving. Our findings suggest that transcriptome conservation in brown algae is associated with cell differentiation stages, but not necessarily linked to embryogenesis. Together with previous work in animals, plants and fungi, we provide further evidence for the generality of a developmental hourglass pattern across complex multicellular eukaryotes.

evolutionary biology↗

bioRxiv2024

Identify compound-protein interaction with knowledge graph embedding of perturbation transcriptomics

The emergence of perturbation transcriptomics provides a new perspective and opportunity for drug discovery, but existing analysis methods suffer from inadequate performance and limited applicability. In this work, we present PertKGE, a method designed to improve compound-protein interaction with knowledge graph embedding of perturbation transcriptomics. PertKGE incorporates diverse regulatory elements and accounts for multi-level regulatory events within biological systems, leading to significant improvements compared to existing baselines in two critical "cold-start" settings: inferring binding targets for new compounds and conducting virtual ligand screening for new targets. We further demonstrate the pivotal role of incorporating multi- level regulatory events in alleviating dataset bias. Notably, it enables the identification of ectonucleotide pyrophosphatase/phosphodiesterase-1 as the target responsible for the unique anti- tumor immunotherapy effect of tankyrase inhibitor K-756, and the discovery of five novel hits targeting the emerging cancer therapeutic target, aldehyde dehydrogenase 1B1, with a remarkable hit rate of 10.2%. These findings highlight the potential of PertKGE to accelerate drug discovery by elucidating mechanisms of action and identifying novel therapeutic compounds.

bioinformatics↗