Biology subjects

Zhou, J.

Publications and source records attributed to Zhou, J..

At least 235 records · Page 13Linked to original sources

bioRxiv2024

BioStructNet: Structure-Based Network with Transfer Learning for Predicting Biocatalyst Functions

Enzyme-substrate interactions are essential to both biological processes and industrial applications. Advanced machine learning techniques have significantly accelerated biocatalysis research, revolutionizing the prediction of biocatalytic activities and facilitating the discovery of novel biocatalysts. However, the limited availability of data for specific enzyme functions, such as conversion efficiency and stereoselectivity, presents challenges for prediction accuracy. In this study, we developed BioStructNet, a structure-based deep learning network that integrates both protein and ligand structural data to capture the complexity of enzyme-substrate interactions. Benchmarking studies with the different algorithms showed the enhanced predictive accuracy of BioStructNet. To further optimize the prediction accuracy for the small dataset, we implemented transfer learning in the framework, training a source model on a large dataset and fine-tuning it on a small, function-specific dataset, using the CalB dataset as a case study. The model performance was validated by comparing the attention heat maps generated by the BioStructNet interaction module, with the enzyme substrate interactions revealed by enzyme-substrate complexes revealed from molecular simulations. BioStructNet would accelerate the discovery of functional enzymes for industrial use, particularly in cases where the training datasets for machine learning are small.

bioinformatics↗

bioRxiv2024

DEMINERS enables clinical metagenomics and comparative transcriptomic analysis by increasing throughput and accuracy of nanopore direct RNA sequencing

Nanopore direct RNA sequencing (DRS) advances RNA biology but is limited by relatively low basecalling accuracy, low throughput, yet high RNA input and costs. Here we introduce a novel DRS toolkit, DEMINERS, which integrates an RNA multiplexing experimental workflow, a machine-learning barcode classifier based on Random Forest and a novel basecaller built on an optimized convolutional neural network providing an additional species-specific training module. With the increased accuracy in barcode classification and basecalling, DEMINERS can demultiplex up to 24 samples and the required RNA input and running time are both substantially reduced. We demonstrated the applications of DEMINERS in clinical metagenomics, cancer transcriptomics and parallel comparison of transcriptomic features in different biological conditions, revealing altered airway microbial diversity in COVID-19 and a potential role of m6A in increasing transcriptomic diversity in glioma and the mature blood-stage of malaria parasites. Overall, DEMINERS is a simple, robust, high-throughput DRS method for accurately estimating transcript levels, poly(A) lengths, and mutation and RNA modification heterogeneity at single-read level, with minimal sequencing biases.

bioinformatics↗

bioRxiv2024

Role of dopamine in reward expectation and predictability during execution of action sequences

BackgroundMesolimbic dopamine (DA) neurons are central to cue guided reward seeking and action sequence learning. Yet, the mechanisms by which cue-induced DA neural activity drives goal-directed or habitual sequence execution remain unknown. MethodsWe designed two novel tasks to isolate the effect of sequence-delineating cues on DA-driven behavioral strategies and learning. In the lever insertion fixed-ratio 5 task (LI5), the lever insertion marked sequence initiation. In the lever retraction fixed-ratio 5 task (LR5), the lever retraction served as both sequence termination and reward-predictive cue. ResultsWe found that sequence initiation and termination cues differentially affect reward expectation during action sequences, with only the termination cue contributing to greater outcome devaluation insensitivity, automaticity and behavioral chunking. Mesolimbic fiber photometry recording revealed that this habit-like behavior was associated with a rapid backpropagation in DA signals from the reward to the immediately preceding cue and with attenuated DA reward prediction error signals, which reflected greater behavioral inflexibility. Finally, in absence of external cues, brief optogenetic stimulation of VTA DA neurons at sequence termination was sufficient to drive automaticity and, to some extent, behavioral chunking. ConclusionOur results highlight the critical role of cue-evoked DA signals at sequence termination in mediating credit assignment and driving the development of habitual action sequence execution.

neuroscience↗

bioRxiv2024

Mammary Epithelial Migration is EMT-Independent

Epithelial tissues are traditionally considered immobile, requiring epithelial-to-mesenchymal transition (EMT) for individual or collective cell migration as mesenchymal cells. However, recent studies indicate that the mammary epithelium migrates via mechanisms distinct from mesenchymal collectives. Here, we demonstrate that mammary epithelium does not undergo EMT before migration, as evidenced by cellular and molecular characteristics, including EMT marker expression at the mRNA, protein, localization, and transcriptomic levels. Contrary to established models, cell-cell adhesion is not diminished but is required for epithelial migration. Furthermore, Snail1 does not repress cell-cell adhesion but instead plays a pivotal role in directing epithelial migration by regulating processes such as the directional movement of individual cells via Cdc42ep5. Our findings reveal the existence of both stationary and migratory states within the epithelium, providing new insights into how epithelial tissues transition between these states, with broad implications for organogenesis and cancer biology.

cell biology↗

bioRxiv2024

P450-Catalyzed Biaryl Macrocyclization of Leaderless Ribosomal Peptides

Macrocyclic peptides containing biaryl motifs, predominantly derived from natural products, are valuable scaffolds due to their structural rigidity and potent bioactivity. However, current synthetic methods remain constrained by the absence of broadly applicable strategies across chemocatalytic and biocatalytic platforms. Here, we discover a versatile P450 enzyme, GpeC, capable of facilitating oxidative C-C/O/N cross-coupling for peptide biaryl macrocyclization (PBC) with leader-independent activity. Crystal structure analysis of GpeC rationalizes leader-independent property and reveals adaptive recognition that enables access to diverse biaryl-linked macrocycles. GpeC exhibits exceptional substrate promiscuity, accommodating 19 of 20 canonical amino acids and diverse noncanonical analogs within a minimal tetrapeptide scaffold. The efficient semi-synthesis of the natural product Rubrin further demonstrated the versatility of GpeC. We further introduced lytic to tetraregion (LTT), a one-step, single enzyme synthesis for modular synthesis of biaryl-cyclized tetrapeptide. Overall, GpeCs robustness and programmability position it as a broadly applicable biocatalyst for the synthesis of biaryl macrocyclic peptides.

biochemistry↗

bioRxiv2024

Interpreting Sleep Activity Through Neural Contrastive Learning

Memories are spontaneously replayed during sleep, a process thought to support memory consolidation. However, capturing this replay in humans has been challenging because unlike wakefulness, sleep EEG is dominated by slow, rhythmic background activity. Moreover, each sleep stage (e.g., NREM, REM) has distinct rhythms, hindering generalisation of models trained on wake-state data. To overcome these challenges, we developed the Sleep Interpreter (SI), a neural network model that decodes memory replay from sleep EEG. In a large dataset comprising 135 participants ([~]1,000 h of overnight sleep; [~]400 h of wake), we employed a TMR-like paradigm with 15 semantically congruent cue-image pairs to tag specific memories. SI was trained separately for NREM and REM using contrastive learning to align neural patterns across wake and sleep, filtering out stage-specific background rhythms. We also examined how slow oscillations and spindle coupling influence decoding in NREM sleep. In a 15-way classification, SI achieved up to 40.02% Top-1 accuracy on unseen subjects. To test generalisability, we followed up with two independent nap experiments in separate samples and applied the trained SI model off-the-shelf. The first probed spontaneous reactivation without auditory cues, while the second used semantic-free sounds with new images. In both, SI successfully decoded reactivation during sleep that correlated with post-nap memory performance. By openly sharing our dataset and SI system, we provide a unique resource for advancing research on memory and learning during sleep, and related disorders.

neuroscience↗

bioRxiv2024

Continual integration of single-cell multimodal data with MIRACLE

Single-cell sequencing technologies have revolutionized our understanding of cellular heterogeneity and facilitated the construction of multi-omics cell atlases via data integration. However, updating these atlases with new data conventionally requires reintegration of all data and is computationally intensive, hindering timely updates and dynamic adjustments in biological and medical research. To address this challenge, we present Multimodal Integration with Continual Learning (MIRACLE), a novel online learning framework for the adaptive and efficient integration of single-cell multimodal data. MIRACLE employs dynamic architectures and data rehearsal strategies to support continual learning, allowing diverse data to be integrated while minimizing information loss over time. Our evaluations demonstrate that MIRACLE achieves accurate online integration with reduced computational requirements, effectively updating and expanding atlases with new cross-tissue and cross-modal data, and precisely identifying novel cell types and transferring labels across datasets. MIRACLE provides an efficient and flexible tool for single-cell community to integrate, share and explore biological knowledge from single-cell multimodal data.

bioinformatics↗

bioRxiv2024

Growth and defence in a unicellular microalga: the coexistence of r and K strategies

r (high reproductivity but stress-sensitive offspring) and K (low reproductivity but high stress-resistant offspring) strategies are two classical life history theories. Contrary to the belief that one species only has a single strategy (i.e., either r or K), we present experimental evidence for the coexistence of both strategies in the isogenic microalga Haematococcus lacustris. Under standard conditions, swimming vegetative cells (SVC) normally grew and divided via binary fission (r strategy) and gradually shifted into non-swimming cysts (NSC) over time. Intriguingly, unlike the prevailing notion that NSC cannot propagate under stress, they were found to reproduce via multiple fission at a barely detectable growth rate, resulting in stress-resistant, non-mobile daughter cells and demonstrating a complete life history (K strategy). Collectively, our findings indicate that H. lacustris adopts both r and K strategies throughout its life history. This research enhances the understanding of the adaptation and survival of microbial populations in changing environments.

ecology↗

bioRxiv2024

Importance of higher-order epistasis in large protein sequence-function relationships

Protein sequence-function relationships are inherently complex, as amino acids at different positions can interact in highly unpredictable ways. A key question for protein evolution and engineering is how often epistasis extends beyond pairwise interactions to involve three or more positions. Although experimental data has accumulated rapidly in recent years, addressing this question remains challenging, as the number of possible interactions is typically enormous even for proteins of moderate size. Here, we introduce an interpretable machine learning framework for studying higher-order epistasis scalable to full-length proteins. Our model builds on the transformer architecture, with key modifications allowing us to assess the importance of higher-order interactions by fitting a series of models with increasing complexity. Applying our method to 10 large protein sequence-function datasets, we found that while additive effects explain the majority of the variance, within the epistatic component, the contribution of higher-order epistasis ranges from negligible to up to 60%. We also found higher-order epistasis is particularly important for generalizing locally sampled fitness data to distant regions of sequence space and for modeling an additional multi-peak fitness landscape. Our findings suggest that higher-order epistasis can play important roles in protein sequence-function relationships, and thus should be properly considered in protein engineering and evolutionary data analysis.

genetics↗

bioRxiv2024

A lymphoid tissue chemokine checkpoint prevents loss of CD8+ T cell functionality

The generation of effector CD8+ T cells (TEFF) requires activation of naive CD8+ T cells (TN) by dendritic cells (DCs) within lymphoid tissue. To date, it remains elusive how the duration of TN-DC interactions and integration of activation signals are controlled in vivo. Here, we report that lymphoid stroma-secreted ligands for CCR7 constrained interaction duration by gradually inducing CD8+ T cell release from DCs. At late time points of interactions, CCR7 ligands repositioned the F-actin-promoting factor DOCK2 away from the DC interface to enable CD8+ T cell detachment, proliferation onset and acquisition of cytotoxicity. Lack of CCR7 signaling, as during ex vivo activation or in chronically inflamed lymphoid tissue, caused sustained T cell-DC interactions, and generated dysfunctional TEFF with high expression of inhibitory receptors, impaired antimicrobial activity, and poor recall responses. In sum, our findings uncover that lymphoid stromal chemokines act as built-in "disruptors" of T cell-DC interactions for long-term preservation of TEFF functionality.

immunology↗

bioRxiv2024

An integrated stress response-independent role of GCN2 prevents excessive ribosome biogenesis and mRNA translation

The Integrated Stress Response (ISR) is a corrective physiological program to restore cellular homeostasis that is based on the attenuation of global protein synthesis and a resource-enhancing transcriptional programme. GCN2 is the oldest of four kinases that are activated by diverse cellular stresses to trigger the ISR and acts as the primary responder to amino acid shortage and ribosome collisions. Here, using a broad multi-omic approach, we uncover an ISR-independent role of GCN2. GCN2 inhibition or depletion in the absence of discernible stress causes excessive protein synthesis and ribosome biogenesis, perturbs the cellular translatome, and results in a dynamic and broad loss of metabolic homeostasis. Cancer cells that rely on GCN2 to keep protein synthesis in check under conditions of full nutrient availability depend on GCN2 for survival and unrestricted tumour growth. Our observations define an ISR-independent role of GCN2 in regulating the cellular proteome and translatome and suggest new avenues for cancer therapies based on unleashing excessive mRNA translation.

cancer biology↗

bioRxiv2024

Deciphering the Cell Specific Transcript Heterogeneity and Alternative Splicing during the Early Embryonic Development of Zebrafish

Cell fate determination during early embryonic development is a complex process modulated by gene expression. The intricate interplay of transcriptional and post-transcriptional regulation is integral to the developmental trajectory of embryogenesis, yet how RNA processing may contribute to early development programming is largely elusive. Leveraging recent technological advances in single-molecule nanopore sequencing, we developed a single-cell long-read transcriptome sequencing technology, allowing a clear view of transcript diversity during zebrafish embryogenesis during pre- and post-zygotic genome activation (ZGA). A closer examination of the dynamic transcript usage and potential alternative splicing revealed that abundant stage-specific transcripts with differential coding potentials are involved in distinct biological functions. Specifically, we identified two cell populations at the onset of ZGA based on isoform diversity instead of gene profiling, which followed divergent developmental trajectories toward the ectoderm and the presumptive ectoderm. These two populations of cells were characterized by divergent splicing regulations linked to differential RNA-binding proteins, including SNRPA and SFPQ. Altogether, using the single-cell long-read transcriptome sequencing strategy, we work has revealed the cell-specific transcriptome dynamics contributing to the cell fate determination during embryogenesis.

developmental biology↗

bioRxiv2024

METTL3 inhibition-activated cGAS/STING axis enhances immunotherapy and PARP inhibitor sensitivity of lung carcinoma

The cGAS/STING-mediated type I interferon response can augment antitumor activity, while the regulatory factors within this innate immune response remain elusive. Herein we found that the RNA m6A methyltransferase METTL3 was upregulated in lung carcinoma tissues. Elevated METTL3 level was correlated with diminished CD8+ T cell infiltration and cancer progression in lung carcinoma patients. METTL3 deficiency exacerbated nuclear DNA leakage into the cytoplasm, activating the cGAS pathway and thereby enhancing anti-tumor immunity. Mechanistically, METTL3 deficiency reduced the homologous recombination repair efficacy via downregulation of MSH5, a mutS family protein involved in DNA mismatch repair, leading to increased cytosolic DNA levels. m6A methylation of A2521 of MSH5 stabilized its mRNA via binding with IGF2BP2. On the other hand, m6A methylation of A1545 at the CDS of cGAS decreased mRNA stability and regulates its protein expression. Functionally, knockdown of METTL3 sensitized lung carcinoma cells to the PARP inhibitors. In vivo and clinical data confirmed the positive roles of METTL3 inhibition-activated cGAS/STING axis in tumor growth and lung adenocarcinoma progression. Collectively, METTL3 inhibition activates the cGAS/STING-mediated anti-tumor immunity via induction of cytosolic DNA and cGAS expression, which in turn regulate PARP inhibitor response and cancer progression in lung carcinoma.

cancer biology↗

bioRxiv2024

A wheat tandem kinase and NLR pair confers resistance to multiple fungal pathogens

Recently discovered tandem kinase proteins (TKPs) are pivotal to the innate immune systems of cereal plants, yet how they initiate plant immune responses remains unclear. This report identifies the wheat protein WTN1, a non-canonical NLR receptor featuring tandem NB-ARC domains, as crucial for WTK3-mediated disease resistance. Both WTK3 and its allelic variant Rwt4, known for conferring resistance to wheat powdery mildew and blast respectively, are capable of recognizing the blast effector PWT4, and activate WTN1 to form calcium-permeable channels, akin to ZAR1 and Sr35. This study unveils a unique plant defense mechanism wherein TKPs and associated NLRs operate as "sensor-executor" pairs against fungal pathogens. Additionally, evolutionary analyses reveal a co-evolutionary trajectory of the TKP-NLR module, highlighting their synergistic role in triggering plant immunity. One Sentence SummaryAn ancient synergistic TKP-NLR pair triggers innate immunity for multiple disease resistance in wheat.

plant biology↗

bioRxiv2024

Single-turnover kinetic analysis of non-LTR retrotransposition defines the mechanism and rate constants governing each step

Site specific retrotransposon-mediated gene therapy has the potential to revolutionize medicine by allowing insertion of large gene cargos. Despite decades of effort, the reaction sequence remains to be fully elucidated limiting the ability to engineer improved activity for gene insertion. Here we provide a kinetic/mechanistic framework for R2 non-LTR retrotransposition. Single turnover measurements and global data fitting defined the rate constants for each step in the pathway involving 1st-strand DNA cleavage to provide a DNA primer, reverse transcription to copy the RNA, 2nd-strand DNA cleavage to provide the second primer, and 2nd-strand synthesis to make duplex cDNA. Sequence analysis of the cDNA confirms accurate replication of the 1400 nt RNA used in this study. This represents the only complete analysis of the reaction sequence and first observation of 2nd-strand synthesis in vitro. We provide a kinetic framework to understand non-LTR retrotransposition, which provide a basis to engineer improved activity.

biochemistry↗

bioRxiv2024

Evolution of the Spatial transcriptomic landscape during the progression of high-grade pancreatic intraductal papillary mucinous neoplasms to invasive cancer

Intraductal papillary mucinous neoplasms (IPMN) is one of the known precancerous lesions. Patients prognosis is aggravated as IPMN transforms into invasive Pancreatic Ductal Adenocarcinoma (PDAC). The molecular mechanisms underlying this progression lack effective experimental models and urgently need to be elaborated. We performed spatial transcriptomics (ST) on fresh tissue samples from the same patient including normal pancreas, high-grade IPMN, and invasive PDAC, and described the step-by-step development of transcriptional landscape including clone evolution and adjacent TME feature variation. Our findings identified the master transcript factors and critical signaling pathways promoting IPMN progression to invasive PDAC. Additionally, both IPMN and PDAC harbored the ELF3, MYC, and KLF4 amplification. The Spatial CNV profile demonstrated significant heterogeneity among PDAC in their spatial distribution compared to IPMN, with seven distinct subclones showing diverse functions, such as hypoxia, oxidative phosphorylation, and epithelial-mesenchymal Transition. We observed a marked shift in the immune landscape, with the depletion of CD4+ T-cells and dendritic cells and an increase in immune-suppressive M2 macrophages in invasive PDAC, indicating a transition to an immune-evasive microenvironment. Additionally, cancer-associated fibroblasts (CAFs), particularly myofibroblastic CAFs, were enriched adjacent to invasive PDAC, suggesting their active role in tumor progression. By leveraging spatial transcriptomic analysis, our study provides comprehensive insights into the intricate molecular landscape that underlies the progression of IPMNs to invasive PDAC. These findings not only enhance our understanding of this complex process but also offer valuable knowledge for early diagnosis and intervention. HighlightsO_LISpatial CNV analysis reveals clonal evolution and distinct subclones in PDAC. C_LIO_LIKey drivers like ELF3, MYC, and KLF4 are amplified in both IPMN and invasive PDAC. C_LIO_LIImmune landscape shifts from pro-inflammatory in IPMN to immune-evasive in PDAC. C_LIO_LIEnrichment of myofibroblastic CAFs suggests their role in tumor progression C_LI

cancer biology↗

bioRxiv2024

ActSort: An active-learning accelerated cell sorting algorithm for large-scale calcium imaging datasets

Recent advances in calcium imaging enable simultaneous recordings of up to a million neurons in behaving animals, producing datasets of unprecedented scales. Although individual neurons and their activity traces can be extracted from these videos with automated algorithms, the results often require human curation to remove false positives, a laborious process called cell sorting. To address this challenge, we introduce ActSort, an active-learning algorithm for sorting large-scale datasets that integrates features engineered by domain experts together with data formats with minimal memory requirements. By strategically bringing outlier cell candidates near the decision boundary up for annotation, ActSort reduces human labor to about 1-3% of cell candidates and improves curation accuracy by mitigating annotator bias. To facilitate the algorithms widespread adoption among experimental neuroscientists, we created a user-friendly software and conducted a first-of-its-kind benchmarking study involving about 160,000 annotations. Our tests validated ActSorts performance across different experimental conditions and datasets from multiple animals. Overall, ActSort addresses a crucial bottleneck in processing large-scale calcium videos of neural activity and thereby facilitates systems neuroscience experiments at previously inaccessible scales.

neuroscience↗

bioRxiv2024

A COVID-19 Rapid Antigen Test Employing Upconversion Nanoparticles

The COVID-19 pandemic has underscored the critical need for rapid and accurate diagnostic tools. Current methods, including PCR and rapid antigen tests (RAT), have limitations in speed, sensitivity, and the requirement for specialized equipment and trained personnel. Nanotechnology, particularly upconversion nanoparticles (UCNPs), offer a promising alternative due to their unique optical properties. UCNPs can convert low-energy near-infrared (NIR) light into higher-energy visible light, making them ideal for use as optical probes in single molecule detection and point of care applications. This study, initiated in early 2020, explores the opportunity of using highly doped UCNPs (40%Yb3+/4%Er3+) in lateral flow assay (LFA) for the early diagnosis of COVID-19. The UCNPs-based LFA testing demonstrated a minimum detection concentration of 100 pg/mL for SARS-CoV-2 antigen and 105 CCID50/mL for inactivated virus. Clinical trials, conducted in Malaysia and Western Australia independently, showed that the technique was at least 100 times more sensitive than commercial RAT kits, with a sensitivity of 100% and specificity of 91.34%. The development process involved multidisciplinary collaborations, resulting in the Virulizer device, an automated strip reader for point-of-care testing. This work sets a reference for future development of highly sensitive and quantitative rapid antigen tests, aiming for the Limits of Detection (LoD) in the range of sub-ng/mL.

bioengineering↗