Biology subjects

Hu, Z.

Publications and source records attributed to Hu, Z..

At least 163 records · Page 9Linked to original sources

bioRxiv2024

CellWalker2: multi-omic discovery of hierarchical cell type relationships and their associations with genomic annotations

CellWalker2 is a graph diffusion-based method for single-cell genomics data integration. It extends the CellWalker model by incorporating hierarchical relationships between cell types, providing estimates of statistical significance, and adding data structures for analyzing multi-omics data so that gene expression and open chromatin can be jointly modeled. Our open-source software enables users to annotate cells using existing ontologies and to probabilistically match cell types between two or more contexts, including across species. CellWalker2 can also map genomic regions to cell ontologies, enabling precise annotation of elements derived from bulk data, such as enhancers, genetic variants, and sequence motifs. Through simulation studies, we show that CellWalker2 performs better than existing methods in cell type annotation and mapping. We then use data from the brain and immune system to demonstrate CellWalker2s ability to discover cell type-specific regulatory programs and both conserved and divergent cell type relationships in complex tissues.

bioinformatics↗

bioRxiv2024

Enhancing Biosecurity with Watermarked Protein Design

The biosecurity issue arises as the capability of deep learning-based protein design has rapidly increased in recent years. To address this problem, we propose a new general framework for adding watermarks to protein sequences designed by various sampling-based deep learning models. Compared to currently proposed protein design regulation procedures, watermarks ensure robust traceability and maintain the privacy of protein sequences. Moreover, using our framework does not decrease the performance or accessibility of the protein design tools.

bioinformatics↗

bioRxiv2024

Multi-omics delineate growth factor network underlying exercise effects in an Alzheimer's mouse model

Physical exercise represents a primary defense against age-related cognitive decline and neurodegenerative disorders like Alzheimers disease (AD). To impartially investigate the underlying mechanisms, we conducted single-nucleus transcriptomic and chromatin accessibility analyses (snRNA-seq and ATAC-seq) on the hippocampus of mice carrying AD-linked NL-G-F mutations in the amyloid precursor protein gene (APPNL-G-F) following prolonged voluntary wheel-running exercise. Our study reveals that exercise mitigates amyloid-induced changes in both transcriptomic expression and chromatin accessibility through cell type-specific transcriptional regulatory networks. These networks converge on the activation of growth factor signaling pathways, particularly the epidermal growth factor receptor (EGFR) and insulin signaling, correlating with an increased proportion of immature dentate granule cells and oligodendrocytes. Notably, the beneficial effects of exercise on neurocognitive functions can be blocked by pharmacological inhibition of EGFR and the downstream phosphoinositide 3-kinases (PI3K). Furthermore, exercise leads to elevated levels of heparin-binding EGF (HB-EGF) in the blood, and intranasal administration of HB-EGF enhances memory function in sedentary APPNL-G-F mice. These findings offer a panoramic delineation of cell type-specific hippocampal transcriptional networks activated by exercise and suggest EGF-related growth factor signaling as a druggable contributor to exercise-induced memory enhancement, thereby suggesting therapeutic avenues for combatting AD-related cognitive decline.

animal behavior and cognition↗

bioRxiv2024

Changes in Cerebrospinal Fluid Proteins across the Spectrum of Untreated and Treated Chronic HIV-1 Infection

Using the Olink Explore 1536 platform, we measured 1,463 unique proteins in 303 cerebrospinal fluid (CSF) specimens from four clinical centers that included uninfected controls and 12 groups of people living with HIV-1 infection representing the spectrum of progressive untreated and treated chronic infection. We present three initial analyses of these measurements: an overview of the CSF protein features of the sample; correlations of the CSF proteins with CSF HIV-1 RNA and neurofilament light chain protein (NfL) concentrations; and comparison of the CSF proteins in HIV-associated dementia (HAD) and neurosymptomatic CSF escape (NSE). These reveal a complex but coherent picture of CSF protein changes that includes highest concentrations of many proteins during CNS injury in the HAD and NSE groups and variable protein changes across the course of neuroasymptomatic systemic HIV-1 progression, including two common patterns, designated as lymphoid and myeloid patterns, related to the principal involvement of their underlying inflammatory cell lineages. Antiretroviral therapy reduced CSF protein perturbations, though not always to control levels. The dataset of these CSF protein measurements, along with background clinical information, is posted online. Extended studies of this unique dataset will provide more detailed characterization of the dynamic impact of HIV-1 infection on the CSF proteome across the spectrum of HIV-1 infection, and further the mechanistic understanding of HIV-1-related CNS pathobiology.

microbiology↗

bioRxiv2024

Metabolic immaturity of newborns and breast milk bile acids are the central determinants of heightened neonatal vulnerability to norovirus diarrhea

Noroviruses are the leading global cause of acute gastroenteritis, responsible for 685 million annual cases. While all age groups are susceptible to noroviruses, children are vulnerable to more severe infections than adults, underscored by 200 million pediatric cases and up to 200,000 deaths in children annually. Understanding the basis for the increased vulnerability of young hosts is critical to developing effective treatments. The pathogenic outcome of any enteric virus infection is governed by a complex interplay between the virus, intestinal microbiota, and host immune factors. A central mediator in these complex relationships are host- and microbiota-derived metabolites. Noroviruses bind a specific class of metabolites, bile acids, which are produced by the host and then modified by commensal bacterial enzymes. Paradoxically, bile acids can have both proviral and antiviral roles during norovirus infections. Considering these opposing effects, the microbiota-regulated balance of the bile acid pool may be a key determinant of the pathogenic outcome of a norovirus infection. The bile acid pool in newborns is unique due to immaturity of host metabolic pathways and developing gut microbiota, which could underlie the vulnerability of these hosts to severe norovirus infections. Supporting this concept, we demonstrate herein that microbiota and their bile acid metabolites protect from severe norovirus diarrhea whereas host-derived bile acids promote disease. Remarkably, we also report that maternal bile acid metabolism determines neonatal susceptibility to norovirus diarrhea during breastfeeding by delivering proviral bile acids to the newborn. Finally, directed targeting of maternal and neonatal bile acid metabolism can protect the neonatal host from norovirus disease. Altogether, these data support the conclusion that metabolic immaturity in newborns and ingestion of proviral maternal metabolites in breast milk are the central determinants of heightened neonatal vulnerability to norovirus disease.

microbiology↗

bioRxiv2024

LETSmix: a spatially informed and learning-based domain adaptation method for cell-type deconvolution in spatial transcriptomics

Spatial transcriptomics (ST) has revolutionized our understanding of gene expression patterns by incorporating spatial context. However, many ST technologies operate on heterogeneous cell mixtures due to limited spatial resolution. To resolve cell type composition at each sequencing spot, several deconvolution methods have been proposed. Yet, these approaches often underutilize spatial context inherent in ST data and paired histopathological images, meanwhile overlooking domain variances between ST and reference single-cell RNA sequencing (scRNA-seq) data. Here, we present LETSmix, a novel deconvolution method that enhances spatial correlations within ST data using a tailored LETS filter, and employs a mixup-augmented domain adaptation strategy to address domain shifts. The performance of LETSmix was validated across diverse ST platforms and tissue types, including 10x Visium human dorsolateral prefrontal cortex, ST human pancreatic ductal adenocarcinoma, 10x Visium mouse liver, and Stereo-seq mouse olfactory bulb datasets. Our findings demonstrate that the proposed method accurately estimates cell type proportions and effectively maps them to the expected regions, establishing a new record among current state-of-the-art models. LETSmix is expected to serve as a robust tool for advancing studies on cellular composition and spatial architecture in spatial transcriptomics. GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=149 SRC="FIGDIR/small/591425v2_ufig1.gif" ALT="Figure 1"> View larger version (35K): org.highwire.dtl.DTLVardef@1b264cborg.highwire.dtl.DTLVardef@f77e8forg.highwire.dtl.DTLVardef@9609ddorg.highwire.dtl.DTLVardef@61af58_HPS_FORMAT_FIGEXP M_FIG C_FIG

genomics↗

bioRxiv2024

An extended omnigenic model explains genome-phenome relationships for complex traits in global sorghum diversity

A central finding of complex trait genetics is the geometric distribution of effect sizes, but the biological basis of this phenomena is not understood. The omnigenic model (OM) could explain this architecture, with oligogenic variation arising from direct regulatory genes (core genes) and polygenic variation from indirect regulators (peripheral trans regulators). Plant yield is a canonical complex trait and here we tested the OM using genome-phenome analysis of biomass yield in global sorghum diversity. We used field-based phenomics to characterize dynamic growth and yield formation traits, then decomposed oligogenic and polygenic components of variation using genome-wide association studies (GWAS), genome-wide prediction (GWP), and tissue-specific transcriptome analyses. We identified major dynamic QTL, including several persistent (Dw1, Dw3) or transient (ELF3) QTL at known genes, consistent with the oligogenic-core of the OM. Next, we evaluated a key prediction on the peripheral-polygenic component - a positive correlation between GWP marker loadings and gene expression in relevant tissues. GWP loadings are indeed correlated with gene expression in relevant tissues. However, correlations are often higher with non-relevant tissues from earlier growth stages or tissues, which is not predicted by within-tissue trans regulation for the peripheral-polygenic component (the "omnigenic in space" model). Therefore, these findings suggest that genes expressed in early growth stages, with indirect effects on later traits, are contributing to polygenic variation ("omnigenic in time"). Together, our findings suggest that an extended OM, with regulatory effects both in space and in time, could explain the ubiquitous geometric genetic architecture of complex traits.

genetics↗

bioRxiv2024

A rat model of cirrhosis with well-differentiated hepatocellular carcinoma induced by thioacetamide

Hepatocellular carcinoma (HCC) is a leading cause of cancer-related deaths, and commonly associated with hepatic fibrosis or cirrhosis. This study aims to establish a rat model mimicking the progression from liver fibrosis to cirrhosis and subsequently to HCC using thioacetamide (TAA). We utilized male Lewis rats, treating them with intra-peritoneal injections of TAA. These rats received bi-weekly injections of either 200 mg/kg TAA or saline (as a control) over a period of 34 weeks. The development of cirrhosis and hepatocarcinogenesis was monitored through histopathological examinations, biochemical markers, and immunohistochemical analyses. Our results demonstrated that chronic TAA administration induced cirrhosis and well-differentiated HCC, characterized by increased fibrosis, altered liver architecture, and enhanced hepatocyte proliferation. Biochemical analyses revealed significant alterations in liver function markers, including elevated alpha-fetoprotein (AFP) levels, without affecting kidney function or causing significant weight loss or mortality in rats. This TAA-induced cirrhosis and HCC rat model successfully replicates the clinical progression of human HCC, including liver function impairment and early-stage liver cancer characteristics. It presents a valuable tool for future research on the mechanisms of antitumor drugs in tumor initiation and development.

pathology↗

bioRxiv2024

Co-occurrence of Direct and Indirect Extracellular Electron Transfer Mechanisms during Electroactive Respiration in a Dissimilatory Sulfate Reducing Bacterium

Extracellular electron transfer (EET) propels microbial fuel cell (MFC) technology and contributes to the mobility of redox active minerals and microbial syntrophy in nature. Sulfate-reducing bacteria (SRB), especially the genus Desulfovibrio corrode metal electrodes but are of interest for sulfate-containing MFCs providing wastewater treatment. Although extensive studies on SRB-mediated metal electrode corrosion have been done, there remain knowledge gaps on SRB EET to electrodes. We aimed to determine SRB EET mechanisms towards improving SRB performance in MFC wastewater treatment. Our MFCs with Desulfovibrio vulgaris Hildenborough (DvH), a model SRB, indicated that DvH can harvest and send electrons to the carbon cloth electrode. Electricity production with a maximum power density of [~]0.074 W/m2 was observed when the ratio of lactate (electron and carbon donor) to sulfate (electron acceptor) was 60:20 and 0:10 in the anodic and cathodic chamber, respectively. Patterns in current production compared to variations of electron donor/acceptor ratios in the anode and cathode suggested that attachment of DvH to the electrode and biofilm density were critical for effective electricity generation. Analysis of DvH biofilms at different conditions (planktonic dissimilatory sulfate reduction respiration vs. electroactive respiration) by electron microscopy indicated DvH utilized filaments that resemble nano-pili to attach on electrodes and facilitate EET from cell-to-cell and to the electrode. Proteomics profiling of electroactive respiration proteins indicated DvH adapted to electroactive respiration by presenting more pili-, flagellar-related proteins and histidine kinases on electrodes. To investigate the role of pili and biofilm, we grew two DvH mutants in MFCs under the same conditions. The mutant with a deletion of the major pilus-producing gene yielded less voltage and far less attachment to the electrode, suggesting the importance of pili in EET. The mutant with a deficiency in biofilm formation, however, did not eliminate current production indicating the existence of indirect EET. Untargeted metabolomics profiling showed flavin-based metabolites, potential electron shuttles, were dysregulated between respiration modes. This work revealed the metabolic flexibility of DvH to thrive in less than ideal conditions with solid surfaces as both an electron acceptor (growth on anode) and donor (growth on cathode) by using a combination of direct and indirect EET mechanisms. Understanding DvH EET mechanism could enhance the application of DvH in MFCs treating wastewater. ImportanceWe explored the application of Desulfovibrio vulgaris Hildenborough in microbial fuel cells (MFC) and investigated its potential extracellular electron transfer (EET) mechanism. We also conducted untargeted proteomics and metabolomics profiling, offering insights into how DvH adapts metabolically to different electron donors and acceptors. An understanding of the EET mechanism and metabolic flexibility of DvH holds promise for future uses including bioremediation or enhancing efficacy in MFCs for wastewater treatment applications.

microbiology↗

bioRxiv2024

Unveiling neoadjuvant chemotherapy-induced immune landscape remodeling and metabolic reprogramming in lung adenocarcinoma by scRNA-sequencing

BackgroundChemotherapy is widely used to treat lung adenocarcinoma (LUAD) patients comprehensively. Considering the limitations of chemotherapy due to drug resistance and other issues, it is crucial to explore the impact of chemotherapy and immunotherapy on these aspects. Methods and MaterialsTumor samples from nine LUAD patients, of which four only received surgery and five received neoadjuvant chemotherapy, were subjected to scRNA-seq analysis. In vitro and in vivo assays, including flow cytometry, immunofluorescence, Seahorse assay, and tumor xenograft models, were carried out to validate our findings. ResultsA total of 83,622 cells were enrolled for subsequent analyses. The composition of cell types exhibited high heterogeneity across different groups. Functional enrichment analysis revealed that chemotherapy drove significant metabolic reprogramming in tumor cells and macrophages. We identified two subtypes of macrophages: Anti-mac cells (CD45+CD11b+CD86+) and Pro-mac cells (CD45+CD11b+ARG+) and sorted them by flow cytometry. The proportion of Pro-mac cells in lung adenocarcinoma tissues increased significantly after neoadjuvant chemotherapy. Pro-mac cells promote tumor growth and angiogenesis and also suppress tumor immunity. Moreover, through analyzing the remodeling of T and B cells induced by neoadjuvant therapy, we noted that chemotherapy ignited a relatively more robust immune cytotoxic response towards tumor cells. ConclusionOur study demonstrates that chemotherapy induces metabolic reprogramming within the TME of LUAD, particularly affecting the function and composition of immune cells such as macrophages and T cells. We believe our findings will offer insight into the mechanisms of drug resistance and provide novel therapeutic targets for LUAD in the future.

cancer biology↗

bioRxiv2024

The evolution of 5'UTR length was shaped by natural selection for its functional impact on gene expression

Transcription start site (TSS) marks the first DNA nucleotide of a gene transcribed into RNA. Accumulating evidence suggests that alternative TSS usage (ATU) is widespread across eukaryotes, in response to environmental changes and tissue-specific needs. However, how ATU is coordinated with changes in gene transcription activity and whether it represents a regulated process or merely transcription initiation errors, remains unclear. To address these questions, we conducted integrative analyses of high-resolution TSS maps and translation efficiency data from multiple eukaryotic organisms. Our results reveal that most ATU events co-occur with differential gene expression and that the direction of ATU is largely consistent with changes in transcription levels. These findings suggest that ATU likely works in concert with transcriptional activity to fine-tune protein production, modulating translation efficiency and mRNA stability through alterations in 5'UTR sequence and structural features. Given its functional importance, the evolution of TSS locations may have been shaped by natural selection, leading to heterogeneous 5'UTR lengths in genes with distinct expression demands. This study offers new insights into the complexity of gene regulation and provides a plausible explanation for the highly variable 5'UTR landscape observed within and among eukaryotic genes.

genomics↗

bioRxiv2024

Molecular basis of inhibition of the amino acid transporter B0AT1 (SLC6A19)

The epithelial neutral amino acid transporter B0AT1 (SLC6A19) is the major transporter for the absorption of neutral amino acids in the intestine and their reabsorption in the kidney. Mouse models have demonstrated that lack of B0AT1 can normalize elevated plasma amino acids in rare disorders of amino acid metabolism such as phenylketonuria and urea-cycle disorders. This discovery has resulted in a quest to identify high-affinity selective inhibitors of B0AT1 for the treatment of disorders of amino acid metabolism. Here we employed a medicinal chemistry approach to improve lead compounds that were identified in a previous high-throughput screen. This effort yielded a group of compounds that inhibited B0AT1 with IC50-values of 31-90 nM. High-resolution cryo-EM structures of B0AT1 in the presence of two compounds from this series identified an allosteric binding site in the vestibule of the transporter. Mechanistically, binding of these inhibitors prevents a substantial movement of TM1 and TM6 that is required for the transporter to make a conformational change from an outward open state to the occluded state.

biochemistry↗

bioRxiv2024

RNA Splicing Junction Landscape Reveals Abundant Tumor-Specific Transcripts in Human Cancer

RNA splicing is a critical process governing gene expression and transcriptomic diversity. Despite its importance, a detailed examination of transcript variation at the splicing junction level remains scarce. Here, we perform a thorough analysis of RNA splicing junctions in 34,775 samples across multiple sample types. We identified 29,051 tumor-specific transcripts (TSTs) in pan-cancer, with a majority of these TSTs being unannotated. Our findings show that TSTs are positively correlated with tumor stemness and linked to unfavorable outcomes in cancer patients. Additionally, TSTs display mutual exclusivity with somatic mutations and are overrepresented in transposable element-derived transcripts possessing oncogenic functions. Importantly, TSTs can generate neoepitopes that bind to MHC class I molecules for immunotherapy. Moreover, TSTs can be detected in blood extracellular vesicles from cancer patients. Our results shed light on the intricacies of RNA splicing and offer promising avenues for cancer diagnosis and therapy. In briefThis study thoroughly analyzed RNA splicing junctions in 34,775 samples and identified 29,051 tumor-specific transcripts (TSTs), which may serve as novel cancer driver genes, neoantigens, and circulating biomarkers. Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=153 SRC="FIGDIR/small/579246v1_ufig1.gif" ALT="Figure 1"> View larger version (35K): org.highwire.dtl.DTLVardef@7e1db2org.highwire.dtl.DTLVardef@1cfbe81org.highwire.dtl.DTLVardef@1e427e1org.highwire.dtl.DTLVardef@1f655f5_HPS_FORMAT_FIGEXP M_FIG C_FIG

cancer biology↗

bioRxiv2024

Feedback loop regulation between viperin and viral hemorrhagic septicemia virus through competing protein degradation pathways

Viperin is an antiviral protein that exhibits a remarkably broad spectrum of antiviral activity. Viperin-like proteins are found all kingdoms of life, suggesting it is an ancient component of the innate immune system. However, viruses have developed strategies to counteract viperins effects. Here, we describe a feedback loop between viperin and viral hemorrhagic septicemia virus (VHSV), a common fish pathogen. We show that Lateolabrax japonicus viperin (Ljviperin) is induced by both IFN-independent and IFN-dependent pathways, with the C-terminal domain of Ljviperin being important for its anti-VHSV activity. Ljviperin exerts an antiviral effect by binding both the nucleoprotein (N) and phosphoprotein (P) of VHSV and induces their degradation through the autophagy pathway, which is an evolutionarily conserved antiviral mechanism. However, counteracting viperins activity, N protein targets and degrades transcription factors that up-regulate Ljviperin expression, interferon regulatory factor (IRF) 1 and IRF9, through ubiquitin-proteasome pathway. Together, our results reveal a previously unknown feedback loop between viperin and virus, providing potential therapeutic targets for VHSV prevention. ImportanceViral hemorrhagic septicaemia (VHS) is a contagious disease caused by the viral hemorrhagic septicaemia virus (VHSV), which poses a threat to over 80 species of marine and freshwater fish. Currently, there are no effective treatments available for this disease. Understanding the mechanisms of VHSV-host interaction is crucial for preventing viral infections. Here, we found that, as an ancient antiviral protein, viperin degrades the N and P proteins of VHSV through the autophagy pathway. Additionally, the N protein also impacts the biological functions of IRF1 and IRF9 through the ubiquitin-proteasome pathway, leading to the suppression of viperin expression. Therefore, the N protein may serve as a potential virulence factor for the development of VHSV vaccines and screening of antiviral drugs. Our research will serve as a valuable reference for the development of strategies to prevent VHSV infections.

immunology↗

bioRxiv2024

Modeling Combination Chemo-Immunotherapy for Heterogeneous Tumors

Hypermutable cancers create opportunities for the development of various immunotherapies, such as immune checkpoint blockade (ICB) therapy. However, emergent studies have revealed that many hypermutated tumors have poor prognosis due to heterogeneous tumor antigen landscapes, yet the underlying mechanisms remain poorly understood. To address this issue, we developed mathematical models to explore the impact of combining chemotherapy and ICB therapy on heterogeneous tumors. Our results uncover how chemotherapy reduces antigenic heterogeneity, creating improved immunological conditions within tumors, which, in turn, enhances the therapeutic effect when combined with ICB. Furthermore, our results show that the recovery of the immune system after chemotherapy is crucial for enhancing the response to chemo-ICB combination therapy. Author summaryThe challenge posed by intratumoral heterogeneity is gaining recognition in the field of cancer treatment. Despite the success of immune checkpoint blockade (ICB) therapies in enhancing overall survival across various cancer types, the complexity of therapeutic responses persists due to the heterogeneity of tumor antigens. In this study, we developed mathematical models to explore the evolutionary dynamics of tumors with both homogeneous and heterogeneous antigen landscapes. Our analysis reveals that tumors with heterogeneity exhibit resistance to ICB therapy, unlike their homogeneous counterparts which respond positively. Additionally, our models demonstrate that early treatment of heterogeneous tumors with chemotherapy leads to significant remission but also rapid recurrence. Notably, we identified a fascinating trade-off associated with chemotherapy--while suppressing the immune system, it creates a tumor immunological environment that becomes more conducive to ICB therapy. Finally, our modeling highlights the augmented response observed in tumors subjected to a chemo-ICB combination and shows the crucial role of immune recovery in the context of combination therapy.

evolutionary biology↗

bioRxiv2024

Control of ciliary transcriptional programs during spermatogenesis by antagonistic transcription factors

Existence of cilia in the last eukaryotic common ancestor (LECA) raises a fundamental question in biology: how the transcriptional regulation of ciliogenesis has evolved? One conceptual answer to this question is by an ancient transcription factor regulating ciliary gene expression in both unicellular and multicellular organisms, but examples of such transcription factors in eukaryotes are lacking. Previously, we showed that an ancient transcription factor XAP5 is required for flagellar assembly in Chlamydomonas. Here, we show that XAP5 and XAP5L are two conserved pairs of antagonistic transcription regulators that control ciliary transcriptional programs during spermatogenesis. Male mice lacking either XAP5 or XAP5L display infertility, as a result of meiotic prophase arrest and sperm flagella malformation, respectively. Mechanistically, XAP5 positively regulates the ciliary gene expression by activating the key regulators including FOXJ1 and RFX families during the early stage of spermatogenesis. In contrast, XAP5L negatively regulates the expression of ciliary genes via repressing these ciliary transcription factors during the spermiogenesis stage. Our results provide new insights into the mechanisms by which temporal and spatial transcription regulators are coordinated to control ciliary transcriptional programs during spermatogenesis.

cell biology↗

bioRxiv2024

Hunting for peptide binders of specific targets with data-centric generative language models

Bioactive peptides have become strong candidates for a variety of clinical therapies due to their diverse advantages, which promotes the development of deep generative models for peptide design. Considering that existing methods cannot effectively deal with the conformational flexibility of peptides and are difficult to capture accurate residue-to-residue interaction dependencies, we propose a unified weakly order-dependent autoregressive language modeling architecture (PepGenWOA) for bioactive peptide generative design with tolerating out-of-order input as inductive bias. The superiority of PepGenWOA is demonstrated by generating three classes of therapeutic peptides, including antimicrobial peptides, anticancer peptides, and peptide binders. For antimicrobial and anticancer peptide generation, PepGenWOA not only comprehensively outperforms state-of-the-art baseline models, but also exhibits a significant propensity to incorporate specific types of residues that are beneficial for antimicrobial or anticancer bioactivity. For characteristic-guided peptide binder generation, the pretrained PepGenWOA is fine-tuned with Mixture-of-Experts-style plugins with lifelong learning paradigm, thereby achieving the best trade-off between memory stability for world knowledge and learning plasticity for new downstream tasks. The subsequent structure-based virtual screening anchors the 7 most promising candidates from the mega-scale synthetic sequence space, reducing the number of candidates for in vitro experimental validation by 5 orders of magnitude. A target binding rate of 28.6% with binding specificity further confirms the effectiveness of the fine-tuning strategy. Overall, PepGenWOA is a unified architecture for peptide generative design that can be flexibly customized for different task requirements, thereby harnessing generative language models to reach the "dark space" of biological language.

bioinformatics↗

bioRxiv2023

Auditory High Entropy Response (A-HER): A Novel Paradigm for Studying Brain Processing of Uncertain Information

This paper introduces a novel experimental paradigm - Auditory High Entropy Response (A-HER), which maximizes the information entropy of auditory stimulus sequences. This allows us to study how the brain processes complex information, rather than isolated individual events. Our analysis of the frequency response of the frontal theta rhythm induced by A-HER indicated a significant increase in signal-to-noise ratio and repeatability compared to zero-entropy Auditory Steady-State Response (A-SSR) and low-entropy mismatch negativity (MMN). We further investigated whether the A-HER response was induced by stimulus sequence differences or uncertainty, and studied its propagation rules. Different principles between evoked and entrained were found in A-HER and A-SSR. In conclusion, the A-HER paradigm, by maximizing stimulus sequence uncertainty, offers a new approach to analyzing how the brain processes uncertain information. It has potential for diagnosing and researching neurological and mental diseases, and for brain-computer interfaces, thus potentially impacting neuroscience, cognitive science, and psychology.

bioengineering↗