Biology subjects

Wu, L.

Publications and source records attributed to Wu, L..

At least 145 records · Page 8Linked to original sources

bioRxiv2024

PBRM1 directs PBAF to pericentromeres and protects centromere integrity

The specialised structure of the centromere is critical for effective chromosome segregation, but its repetitive nature makes it vulnerable to rearrangements. Centromere fragility can drive tumorigenesis, but protective mechanisms preventing fragility are still not fully understood. The PBAF chromatin remodelling complex is frequently misregulated in cancer, but its role in cancer is still not fully characterized. Here, we identify PBAF as a protector of centromere and pericentromere structure with profound consequences for genome stability. A conserved feature of isogenic cell lines lacking PBRM1, a subunit of PBAF, is compromised centromere and pericentromere integrity. PBAF is present at these regions, and the binding pattern changes when PBRM1 is absent. PBRM1 loss creates a dependence on the spindle assembly checkpoint, which represents a therapeutic vulnerability. Importantly, we find that even in the absence of any perturbations, PBRM1 loss leads to centromere fragility, thus identifying a new player in centromere protection.

molecular biology↗

bioRxiv2024

Cortical inhibitory parvalbumin interneurons exhibit metabolic specializations coordinated by PGC-1α that are lost in rodents and humans after traumatic brain injury

Parvalbumin-positive interneurons (PV-INs) regulate neuronal and circuit activity, and their dysfunction is observed across neurological conditions, including traumatic brain injury (TBI), epilepsy, Alzheimers disease, and schizophrenia. PV-INs are particularly vulnerable to cell loss, potentially due to their increased metabolic demands arising from their uniquely high level of electrical activity, which render them susceptible to metabolic pressure. Here, we use single-nucleus RNA-sequencing (snRNAseq) data from a rodent model of TBI, as well as human TBI data, and demonstrate PV-INs have unique metabolic specializations that are lost after injury and can be rescued by in vivo treatment with the glycolytic inhibitor, 2-deoxyglucose. We generated a novel PV-IN transcriptional identity module comprised primarily of genes encoding specialized ion channels, metabolic enzymes, and synaptic machinery, that identifies heterogenous subsets of injury-associated PV-INs with loss of PV-IN transcriptional identity. We show that changes in metabolic specialization are coupled to changes in transcriptional identity in PV-INs and implicate the PV-IN-enriched transcriptional co-activator, Ppargc1a, as a key driver of PV-IN transcriptional metabolic dysfunction. We also identify a family of long non-coding RNAs enriched in this subset of transcriptionally dysfunctional PV-INs that negatively correlates with PV-IN metabolic specialization. Lastly, we utilize these tools to interrogate a published human TBI snRNAseq data set and find nearly identical changes, underscoring the importance of PV-IN metabolic dysfunction in the pathology of TBI.

neuroscience↗

bioRxiv2024

Identify Regulatory eQTLs by Multiome Sequencing in Prostate Single Cells

While genome-wide association studies and expression quantitative trait loci (eQTL) analysis have made significant progress in identifying noncoding variants associated with prostate cancer risk and bulk tissue transcriptome changes, the regulatory effect of these genetic elements on gene expression remains largely unknown. Recent developments in single-cell sequencing have made it possible to perform ATAC-seq and RNA-seq profiling simultaneously to capture functional associations between chromatin accessibility and gene expression. In this study, we tested our hypothesis that this multiome single-cell approach allows for mapping regulatory elements and their target genes at prostate cancer risk loci. We applied a 10X Multiome ATAC + Gene Expression platform to encapsulate Tn5 transposase-tagged nuclei from multiple prostate cell lines for a total of 65,501 high quality single cells from RWPE1, RWPE2, PrEC, BPH1, DU145, PC3, 22Rv1 and LNCaP cell lines. To address data sparsity commonly seen in the single-cell sequencing, we performed targeted sequencing to enrich sequencing data at prostate cancer risk loci involving 2,730 candidate germline variants and 273 associated genes. Although not increasing the number of captured cells, the targeted multiome data did improve eQTL gene expression abundance by about 20% and chromatin accessibility abundance by about 5%. Based on this multiomic profiling, we further associated RNA expression alterations with chromatin accessibility of germline variants at single cell levels. Cross validation analysis showed high overlaps between the multiome associations and the bulk eQTL findings from GTEx prostate cohort. We found that about 20% of GTEx eQTLs were covered within the significant multiome associations (p-value [≤] 0.05, gene abundance percentage [≥] 5%), and roughly 10% of the multiome associations could be identified by significant GTEx eQTLs. We also analyzed accessible regions with available heterozygous SNP reads and observed more frequent association in genomic regions with allelically accessible variants (p = 0.0055). Among these findings were previously reported regulatory variants including rs60464856-RUVBL1 (multiome p-value = 0.0099 in BPH1) and rs7247241-SPINT2 (multiome p-value = 0.0002- 0.0004 in 22Rv1). We also functionally validated a new regulatory SNP and its target gene rs2474694-VPS53 (multiome p-value = 0.00956 in BPH1 and 0.00625 in DU145) by reporter assay and SILAC proteomics sequencing. Taken together, our data demonstrated the feasibility of the multiome single-cell approach for identifying regulatory SNPs and their regulated genes.

genetics↗

bioRxiv2024

CLCC1 Governs Bilayer Equilibration at the Endoplasmic Reticulum to Maintain Cellular and Systemic Lipid Homeostasis

The intricate orchestration of lipid production, storage, and mobilization is vital for cellular and systemic homeostasis1,2. Dysfunctional plasma lipid control represents the major risk factor for cardio-metabolic diseases, the leading cause of human mortality3,4. Within the cellular landscape, the endoplasmic reticulum (ER) is the central hub of lipid synthesis and secretion, particularly in metabolically active hepatocytes in the liver or enterocytes in the gut5,6. Initially assembled in the ER lumen, lipid-ferrying lipoproteins necessitate the cross-membrane transfer of both neutral and phospho-lipids onto the lumenal apolipoprotein B (APOB), in a poorly-defined process7-10. Here we show that trans-bilayer equilibration of phospholipids, regulated by the ER protein CLCC1, determines lipid partition across the ER membrane and consequently systemic lipid homeostasis. CLCC1 partners with the phospholipid scramblase TMEM41B11,12 to recognize imbalanced bilayers and promote lipid scrambling, thereby licensing lipoprotein biogenesis and the subsequent bulk lipid transport. Strikingly, loss of CLCC1 or TMEM41B leads to the emergence of giant lumenal lipid droplets enclosed by extensively imbalanced ER bilayers, and consequently drastically accelerated pathogenesis of metabolic-dysfunction-associated liver steatohepatitis (MASH). The above results establish phospholipid scrambling at the ER as the lynchpin to maintain a dynamic equilibrium. Considering the requirement of trans-bilayer phospholipid equilibration in numerous biological processes, ranging from catabolic autophagy to viral infection13-16, our study may enable further elucidation of a previously under-appreciated homeostatic control mechanism intrinsic to the ER function in lipid biogenesis and distribution.

cell biology↗

bioRxiv2024

Soybean PHR1-regulated low phosphorus-responsive GmRALF22 increases uptake of phosphate via stimulating GmPTs expression

Phosphorus (P) is one of essential macronutrients for plant growth and development. Rapid Alkalization Factors (RALFs) play crucial roles in plant responses to nutrient stresses, however, the functions of Glycine max RALFs (GmRALFs) under low P (LP) stress remain elusive. In this study, we first identified 27 GmRALFs in soybean, then we revealed that GmRALF10, GmRALF11, and GmRALF22 are induced in both roots and leaves, while only GmRALF5, GmRALF6, and GmRALF25 are up-regulated in leaves in LP conditions. Furthermore, GmRALF22 was found to be the target gene of transcription factor GmPHR1, which binds the P1BS cis-element in the promoter of GmRALF22 via electrophoretic mobility shift assay (EMSA) and DUAL-LUC experiment. Colonization of Bacillus subtilis that deliver GmRALF22 increases the expression of high affinity phosphate (Pi) transporter gene GmPT2, GmPT11, GmPT13 and GmPT14, thus increases the total amount of dry matter and soluble Pi in soybean. RNA-sequencing uncovered that GmRALF22 alleviates LP stress by regulating the expression of JA-, SA- and immune-related genes. Finally, we verified that GmRALF22 is dependent on FERONIA to promote Arabidopsis primary root growth under LP conditions. In summary, GmPHR1-GmRALF22 module positively regulates soybeans tolerance to LP. HighlightsO_LISoybean genome has 27 GmRALFs. GmRALF5, GmRALF6, GmRALF10, GmRALF11, GmRALF22 and GmRALF25 are induced in low phosphorus (LP). C_LIO_LIGmPHR1 directly regulate the transcription of GmRALF22 via binding the promoter P1BS cis-element. C_LIO_LISecretion of GmRALF22 protein by Bacillus subtilis promoted soybean growth under LP conditions by improving soybean P nutrition through increased expression of high affinity phosphate (Pi) transporter gene. C_LIO_LIGmRALF22 regulates soybean P nutrition by JA-, SA- and immune-related genes expression at transcriptome level. That application of GmRALF22 promoted primary root growth in Arabidopsis in LP is dependent on FER receptor. C_LI

molecular biology↗

bioRxiv2024

GeoAB: Towards Realistic Antibody Design and Reliable Affinity Maturation

Increasing works for antibody design are emerging to generate sequences and structures in Complementarity Determining Regions (CDRs), but problems still exist. We focus on two of them: (i) authenticity of the generated structure and (ii) rationality of the affinity maturation, and propose GEOAB as a solution. In specific, GeoABDesigner generates CDR structures with realistic internal geometries, composed of a generative geometry initializer (Geo-Initializer) and a position refiner (Geo-Refiner); GeoAB-Optimizer achieves affinity maturation by accurately predicting both the mutation effects and structures of mutant antibodies with the same network architecture as Geo-Refiner. Experiments show that GEOAB achieves state-of-the-art performance in CDR co-design and mutation effect predictions, and fulfills the discussed tasks effectively.

bioinformatics↗

bioRxiv2024

Unbiased metastatic niche-labeling identifies estrogen receptor-positive macrophages as a barrier of T cell infiltration during bone colonization

Microenvironment niches determine cellular fates of metastatic cancer cells. However, robust and unbiased approaches to identify niche components and their molecular profiles are lacking. We established Sortase A-Based Microenvironment Niche Tagging (SAMENT), which selectively labels cells encountered by cancer cells during metastatic colonization. SAMENT was applied to multiple cancer models colonizing the same organ and the same cancer to different organs. Common metastatic niche features include macrophage enrichment and T cell depletion. Macrophage niches are phenotypically diverse between different organs. In bone, macrophages express the estrogen receptor alpha (ER) and exhibit active ER signaling in male and female hosts. Conditional knockout of Esr1 in macrophages significantly retarded bone colonization by allowing T cell infiltration. ER expression was also discovered in human bone metastases of both genders. Collectively, we identified a unique population of ER+ macrophages in the metastatic niche and functionally tied ER signaling in macrophages to T cell exclusion during metastatic colonization. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=165 SRC="FIGDIR/small/593016v1_ufig1.gif" ALT="Figure 1"> View larger version (60K): org.highwire.dtl.DTLVardef@1cd65b6org.highwire.dtl.DTLVardef@1d099e2org.highwire.dtl.DTLVardef@26cd91org.highwire.dtl.DTLVardef@181f5f3_HPS_FORMAT_FIGEXP M_FIG C_FIG HIGHLIGHTSO_LISAMENT is a robust metastatic niche-labeling approach amenable to single-cell omics. C_LIO_LIMetastatic niches are typically enriched with macrophages and depleted of T cells. C_LIO_LIDirect interaction with cancer cells induces ER expression in niche macrophages. C_LIO_LIKnockout of Esr1 in macrophages allows T cell infiltration and retards bone colonization. C_LI

cancer biology↗

bioRxiv2024

Single cell profiling of bone metastasis ecosystems from multiple cancer types reveals convergent and divergent mechanisms of bone colonization.

Bone represents congenial soil for metastatic seeds and is frequently affected by metastasis of multiple cancer types. The histological and molecular characteristics of bone metastases (BMs) are diverse but poorly understood. Herein, we performed single-cell RNA-seq on 34 BMs from 6 cancer types and identified 3 ecosystem archetypes characterized by enrichment of macrophages/osteoclasts (M{varphi}-OC), regulatory/exhausted T cells (Treg-Tex), and monocytes (Mono), respectively. Breast cancer BMs are mostly the M{varphi}-OC archetype driven by the osteolytic vicious cycle, whereas kidney cancers BMs mainly belong to the Treg-Tex archetype that lacks osteoclasts. Lung cancers BMs evenly distributed across all archetypes. Further analyses revealed parallel mechanisms of immunosuppression and bone remodeling. Elevated estrogen signaling distinguishes macrophages in the M{varphi}-OC subtype, which was investigated in a companion study. Together, we elucidated that divergent mechanisms toward bone colonization and that BMs of different origins can adopt the same mechanism through convergent evolution or adaptation. HIGHLIGHTSO_LIAnalyses of bone metastases from 6 cancer types revealed three immune archetypes. C_LIO_LIArchetypes diverge on immune trajectories, and features of tumor and stromal cells. C_LIO_LIDominant cell type in each archetype undergoes convergent evolution. C_LIO_LIRegulatory networks converge on osteoclasts and Tregs to drive archetype formation. C_LI Graphic Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=166 SRC="FIGDIR/small/593027v1_ufig1.gif" ALT="Figure 1"> View larger version (61K): org.highwire.dtl.DTLVardef@82cef6org.highwire.dtl.DTLVardef@1e1b021org.highwire.dtl.DTLVardef@1f2881eorg.highwire.dtl.DTLVardef@1c6874a_HPS_FORMAT_FIGEXP M_FIG C_FIG

cancer biology↗

bioRxiv2024

Histone deacetylase 1 and 2 play essential roles in alveolar macrophage homeostasis

Alveolar macrophages (AMs) are the primary lung-resident macrophages that play a pivotal role in pathogen clearance and surfactant homeostasis. Although the molecular regulation of AM development at the transcriptional level has been more widely studied, the epigenetic mechanisms that maintain AM homeostasis remain incompletely understood. Here, we demonstrate that histone deacetylases HDAC1 and HDAC2 play essential roles in preserving AM integrity. Dual deletion of HDAC1 and HDAC2 resulted in severe defect in AM development, leading to pulmonary alveolar proteinosis (PAP) and increased susceptibility to influenza infection in mice. Mechanistically, HDAC1 and HDAC2 act through their deacetylase activity to suppress the expression of pro-apoptotic protein Bim, thereby enhancing cell survival. Furthermore, HDAC1 and HDAC2 form a functional complex with the histone demethylase LSD1 (KDM1A) to co-regulate a transcriptional program governing AM differentiation. Our findings identify HDAC1 and HDAC2 as central epigenetic regulators essential for alveolar macrophage homeostasis.

immunology↗

bioRxiv2024

Interferon-Inducible Guanylate-Binding Protein 5 Inhibits Replication of Multiple Viruses by Binding to the Oligosaccharyltransferase Complex and Inhibiting Glycoprotein Maturation

Viral infection induces production of type I interferons and expression of interferon-stimulated genes (ISGs) that play key roles in inhibiting viral infection. Here, we show that the ISG guanylate-binding protein 5 (GBP5) inhibits N-linked glycosylation of key proteins in multiple viruses, including SARS-CoV-2 spike protein. GBP5 binds to accessory subunits of the host oligosaccharyltransferase (OST) complex and blocks its interaction with the spike protein, which results in misfolding and retention of spike protein in the endoplasmic reticulum likely due to decreased N-glycan transfer, and reduces the assembly and release of infectious virions. Consistent with these observations, pharmacological inhibition of the OST complex with NGI-1 potently inhibits glycosylation of other viral proteins, including MERS-CoV spike protein, HIV-1 gp160, and IAV hemagglutinin, and prevents the production of infectious virions. Our results identify a novel strategy by which ISGs restrict virus infection and provide a rationale for targeting glycosylation as a broad antiviral therapeutic strategy. HighlightsO_LIThe interferon-stimulated gene GBP5 is induced by SARS-CoV-2 infection in vitro and in vivo. C_LIO_LIER-localized GBP5 restricts N-linked glycosylation of SARS-CoV-2 spike protein, leading to protein misfolding and preventing transport to the Golgi apparatus. C_LIO_LIGBP5 binds to OST complex accessory proteins and potentially blocks access of the catalytic subunit to the spike protein. C_LIO_LIGBP5 inhibits N-glycosylation of key proteins in multiple viruses, including SARS-CoV-2 C_LIO_LIPharmacological inhibition of OST blocks host cell infection by SARS-CoV-2, variants of concern, HIV-1, and IAV. C_LI SignificanceViral infection induces production of type I interferons and expression of interferon-stimulated genes (ISGs) that play key roles in inhibiting viral infection. We found that the interferon-stimulated gene GBP5 is induced by SARS-CoV-2 infection in vitro and in vivo. GBP5 inhibits N-glycosylation of key proteins in multiple viruses, including SARS-CoV-2. Importantly, pharmacological inhibition of Oligosaccharyltransferase (OST) Complex blocks host cell infection by SARS-CoV-2, variants of concern, HIV-1, and IAV, indicating future translational application of our findings.

microbiology↗

bioRxiv2024

A Quantitative Metric of Confidence For Segmentation of Nuclei in Large Spatially Variable Image Volumes

Nuclei segmentation is an important step for quantitative analysis of fluorescence microscopy images. A large volume generally has many different regions containing nuclei with varying spatial characteristics. Automatically identifying nuclei that are challenging to segment can speed up the analysis of biological tissues. Here we show a segmentation technique that provides a metric of segmentation "confidence" for each segmented object in an image volume. This confidence metric can be used either to generate a "confidence map" for visual distinction of reliable from unreliable regions, or in the data space to identify questionable measurements that can be analyzed separately or eliminated from analysis. In an analysis of nuclei in a 3-dimensional image volume, we show that the confidence map correlates well with visual evaluations of segmentation quality, and that the confidence metric correlates well with F1 scores within subregions of the image volume. In addition, we also describe three visualization methods that can visualize the segmentation differences between a segmented volume and a reference volume.

cell biology↗

bioRxiv2024

Unlocking TAS2R14 activation through intricate multi-ligand binding networks

Bitter taste receptors, particularly TAS2R14, play central roles in discerning a wide array of bitter substances, ranging from dietary components to pharmaceutical agents1-3. In addition, TAS2R14 has broad expression in non-gustatory tissues, suggesting its important roles in selective physiological processes and therapeutic potential4. Here, we present cryo-electron microcopy structures of TAS2R14 in complex with flufenamic acid and aristolochic acid, coupling with G protein subtypes gustducin and Gi. Distinct from most known GPCRs, agonists of TAS2R14 bind to multiple intracellular pockets. We highlight the cholesterol molecules occupying an upper transmembrane site, typically the orthosteric pocket in other GPCRs, also binding in the entrance to an intracellular agonist binding pocket, suggesting an endogenous modulatory function5. The structural and mutagenesis analysis illuminate the receptors broad-spectrum ligand recognition and activation via intricate multiple ligand-binding sites. Furthermore, we unveil the structural configuration of gustducin and its interaction with TAS2R14, as well as the coupling mode of Gi1. This investigation should be instrumental in advancing our knowledge of the mechanisms underlying bitter taste recognition and response, with broader relevance to sensory biology and pharmacology.

biophysics↗

bioRxiv2024

TAK1-mediated phosphorylation of PLCE1 represses PIP2 hydrolysis to impede esophageal squamous cancer metastasis

TAK1 is a serine/threonine protein kinase that is a key regulator in a wide variety of cellular processes. However, the functions and mechanisms involved in cancer metastasis are still not well understood. Here, we found that TAK1 knockdown promoted esophageal squamous cancer cell (ESCC) migration and invasion, whereas TAK1 overexpression resulted in the opposite outcome. These in vitro findings were recapitulated in vivo in a xenograft metastatic mouse model. Mechanistically, co-immunoprecipitation and mass spectrometry demonstrated that TAK1 interacted with phospholipase C epsilon 1 (PLCE1), and phosphorylated PLCE1 at serine 1060 (S1060). Functional studies revealed that phosphorylation at S1060 in PLCE1 resulted in decreased enzyme activity, leading to the repression of PIP2 hydrolysis. As a result, the degradation products of PIP2 including diacylglycerol (DAG) and inositol IP3 were reduced, which thereby suppressed signal transduction in the axis of PKC/GSK-3{beta}/{beta}-Catenin. Consequently, expression of cancer metastasis-related genes was impeded by TAK1. Overall, our data indicate that TAK1 plays a negative role in ESCC metastasis, which depends on the TAK1 induced phosphorylation of PLCE1 at S1060.

cancer biology↗

bioRxiv2024

Integrative spatial omics reveals distinct tumor-promoting multicellular niches and immunosuppressive mechanisms in African American and European American patients with TNBC

Racial disparities in the clinical outcomes of triple-negative breast cancer (TNBC) have been well-documented, but the underlying biological mechanisms remain poorly understood. To investigate these disparities, we employed a multi-omic approach integrating imaging mass cytometry and spatial transcriptomics to characterize the tumor microenvironment (TME) in self-identified Black American (BA) and White American (WA) TNBC patients. Our analysis revealed that the TME in BA patients is marked by a network of endothelial cells, macrophages, and mesenchymal-like cells, which correlates with reduced patient survival. In contrast, the WA TNBC microenvironment is enriched in T-cells and neutrophils, indicative of T-cell exhaustion and suppressed immune responses. Ligand-receptor and pathway analyses further demonstrated that BA TNBC tumors exhibit a relatively "immune-cold" profile, while WA TNBC tumors display features of an "inflamed" TME, suggesting the evolution of a unique immunosuppressive mechanism. These findings provide insight into racially distinct tumor-promoting and immunosuppressive microenvironments, which may contribute to the observed differences in clinical outcomes among BA and WA TNBC patients. Statement of SignificanceThis study identifies distinct tumor microenvironment (TME) profiles in Black and White American TNBC patients, providing new insights into the biological mechanisms driving outcome disparities. Our findings highlight the role of the tumor-endothelial-macrophage niche in these disparities, offering a potential therapeutic target for race-inclusive strategies aimed at improving clinical outcomes. By revealing racial differences in treatment response profiles, this work underscores the necessity for tailored therapies in TNBC. These insights lay the groundwork for the development of inclusive, precision-driven treatment approaches that may help mitigate racial disparities and enhance patient outcomes.

cancer biology↗

bioRxiv2024

de Novo Sequencing of Antibodies for Identification of Neutralizing Antibodies in Human Plasma Post SARS-CoV-2 Vaccination

We present a method for sequencing polyclonal IgG enriched from human plasma, employing a combination of de novo sequencing, proteomics, bioinformatics, protein separation, sequencing, and peptide separations. Our study analyzes a single patients IgG antibody response triggered by the Moderna Spikevax mRNA COVID-19 vaccine. From the sequencing data of the natural polyclonal response to vaccination, we generated 12 recombinant antibodies. Six derived recombinant antibodies, including four generated with de novo sequencing, exhibited similar or higher binding affinities than the original natural polyclonal antibody. Our neutralization tests revealed that the six antibodies possess neutralizing capabilities against the target antigen. This research provides insights into sequencing polyclonal IgG antibodies while highlighting the effectiveness and potential of our approach in generating recombinant antibodies with robust binding affinity and neutralization capabilities. Our proposed approach is an advancement in characterizing the IgG response by directly investigating the circulating pool of IgG without relying exclusively on the B-cell repertoire or population. This is crucial as the B-cell analysis may not accurately represent the circulating antibodies. Interestingly, a large proportion (80 to 90%) of the human antibody sequences generated against SARS-CoV-2 in the literature have been derived solely from B-cell analysis. Therefore, the ability to offer a different perspective is crucial in gaining a comprehensive understanding of the IgG response. Significance StatementWe investigate human IgG targeting the receptor binding domain using de novo proteomics. The peripheral B-cell repertoire may not adequately cover all the circulating IgG for human IgG sequencing. Our approach overcomes this limitation by using a de novo protein sequencing on top of standard proteomics. We obtained distinct de novo sequences, showcasing our methods potential. The recombinant proteins we generate possess traits comparable to or surpassing the naturally occurring polyclonal antibodies (pAbs). This study highlights similarities and differences between IgG populations in blood and circulating B-cells, which is crucial for future biologics development.

immunology↗

bioRxiv2024

Effectiveness and efficiency: label-aware hierarchical subgraph learning for protein-protein interaction (laruGL-PPI)

Protein-protein interactions (PPIs) are crucial in various biological processes and their study has significant implications for drug development and disease diagnosis. Existing deep learning methods for PPIs prediction, including graph neural networks (GNNs), have been widely employed as the solutions, while they usually suffer from performance degradation under complex real-world scenarios. We claim that the topological shortcut is one of the key problems contributing negatively to the performance, according to our analysis. By modeling the PPIs as a graph with protein as nodes and interactions as edge types, the prevailing models tend to learn the pattern of nodes degrees rather than intrinsic sequence-structure profiles, leading to the problem termed topological shortcut. In addition, with the emergence of high-throughput experimental methods such as mass spectrometry and protein chip technology, the amount of available PPI data is exploding. The huge data growth leads to intensive computational costs and challenges computing devices, causing infeasibility in practice. To address the discussed problems, we propose a label-aware hierarchical subgraph learning method (laruGL-PPI) that can effectively infer PPIs while being both interpretable and generalizable. Specifically, we introduced edge-based subgraph sampling to effectively alleviate the problems of topological shortcuts and high computing costs. Besides, the inner-outer connections of PPIs are modeled as a hierarchical graph, together with the dependencies between interaction types constructed by a label graph. Extensive experiments on PPI datasets of different scales demonstrate that laruGL-PPI outperforms state-of-the-art PPI prediction methods, particularly in the testing of unseen proteins. Also, our model can recognize crucial sites of proteins, such as surface sites for binding and active sites for catalysis.

bioinformatics↗

bioRxiv2024

PPFlow: Target-aware Peptide Design with Torsional Flow Matching

Therapeutic peptides have proven to have great pharmaceutical value and potential in recent decades. However, methods of AI-assisted peptide drug discovery are not fully explored. To fill the gap, we propose a target-aware peptide design method called PPFO_SCPLOWLOWC_SCPLOW, based on conditional flow matching on torus manifolds, to model the internal geometries of torsion angles for the peptide structure design. Besides, we establish a protein-peptide binding dataset named PPBench2024 to fill the void of massive data for the task of structure-based peptide drug design and to allow the training of deep learning methods. Extensive experiments show that PPFO_SCPLOWLOWC_SCPLOW reaches state-of-the-art performance in tasks of peptide drug generation and optimization in comparison with baseline models, and can be generalized to other tasks including docking and side-chain packing.

molecular biology↗

bioRxiv2024

Target-aware Molecule Generation for Drug Design Using a Chemical Language Model

Generative drug design facilitates the creation of compounds effective against pathogenic target proteins. This opens up the potential to discover novel compounds within the vast chemical space and fosters the development of innovative therapeutic strategies. However, the practicality of generated molecules is often limited, as many designs focus on a narrow set of drug-related properties, failing to improve the success rate of subsequent drug discovery process. To overcome these challenges, we develop TamGen, a method that employs a GPT-like chemical language model and enables target-aware molecule generation and compound refinement. We demonstrate that the compounds generated by TamGen have improved molecular quality and viability. Additionally, we have integrated TamGen into a drug discovery pipeline and identified 7 compounds showing compelling inhibitory activity against the Tuberculosis ClpP protease, with the most effective compound exhibiting a half maximal inhibitory concentration (IC50) of 1.9 M. Our findings underscore the practical potential and real-world applicability of generative drug design approaches, paving the way for future advancements in the field.

biochemistry↗