Biology subjects

He, B.

Publications and source records attributed to He, B..

At least 55 records · Page 3Linked to original sources

bioRxiv2025

EEG-Informed fMRI Analysis Reveals Neurovascular Coupling in Motor Execution and Imagery

The complementary strengths of electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) have driven extensive research into integrating these two noninvasive modalities to better understand the neural mechanisms underlying cognitive, sensory, and motor functions. However, the precise neural patterns associated with motor functions, especially imagined movements, remain unclear. Specifically, the correlations between electrophysiological responses and hemodynamic activations during executed and imagined movements have not been fully elucidated at a whole-brain level. In this study, we employed an EEG-informed fMRI approach on concurrent EEG-fMRI data to map hemodynamic changes associated with dynamic EEG temporal features during motor-related brain activities. We localized and differentiated the hemodynamic activations corresponding to continuous EEG temporal dynamics across multiple motor execution and imagery tasks. Validation against conventional block fMRI analysis demonstrated high precision in identifying regions specific to motor activities, underscoring the accuracy of the EEG-driven model. Beyond the expected sensorimotor activations, the integrated analysis revealed supplementary coactivated regions showing significant negative covariation between blood oxygenation level-dependent (BOLD) activities and sensorimotor EEG alpha power, including the cerebellum and insular cortex. These findings confirmed both the colocalization of EEG and fMRI activities in sensorimotor regions and a negative covariation between EEG alpha band power and BOLD changes. Moreover, the results provide novel insights into neurovascular coupling during motor execution and imagery on a brain-wide scale, advancing our understanding of the neural dynamics underlying motor functions. HighlightsO_LIMovement-related EEG alpha activity negatively covaries with BOLD responses C_LIO_LIEEG-informed fMRI maps the neurovascular correlates across the entire brain C_LIO_LIA unified model disentangles multiple motor-related brain activations C_LIO_LIEEG-informed fMRI results highly overlap with conventional block fMRI results C_LI

neuroscience↗

bioRxiv2024

MethylQUEEN: A Methylation Encoded DNA Foundation Model

DNA 5-methylcytosine (5mC) modification plays a pivotal role in many biological processes, yet 5mC information and pattern hidden behind remains to be explored. Here, we develop Methylation Language Model based on Quintuple Bidirectional Transformer (MethylQUEEN), a novel pre-trained DNA methylation foundation model capable of sensing methylation states and covering the genome-wide methylation landscape. Through tailored methylation-prone pre-training, MethylQUEEN effectively captured epigenetics information hidden within the DNA sequences: it accurately traces DNAs tissue-of-origin, and successfully recovers the expression profile through methylation states. Integrative analysis on MethylQUEENs attention scores also enables us to reveal the unique methylation status of a tissue for precise disease detection, and identifying key regulatory 5mC sites for disease intervention. As a result, MethylQUEEN signifies a new paradigm in methylation analysis for various biological problems. Besides, our study demonstrates the effectiveness of directly integrating methylation information into pre-training, offering new perspectives and methodologies for a range of methylation-related biological processes. It serves as an initial exploration for the development of more comprehensive epigenomic models.

genomics↗

bioRxiv2024

Metformin inhibits nuclear egress of chromatin fragments in senescence and aging

Chronic inflammation is a hallmark of aging and contributes to many age-associated diseases. Metabolic intervention is a strategy to modulate inflammation. However, the connection between inflammation and metabolism during aging remains poorly understood. A mechanism driving chronic inflammation involves cytoplasmic chromatin fragments (CCFs), which appear in senescent cells and aged tissues, activating the cGAS-STING pathway. The size of the CCFs exceeds the capacity of the nuclear pore complex, raising the question of how chromatin fragments enter the cytoplasm. Here, we report that chromatin fragments exit the nucleus via nuclear egress, a membrane trafficking process at the nuclear envelope that shuttles large complexes from the nucleus to the cytoplasm. Inactivating critical nuclear egress ESCRT-III or Torsin proteins results in accumulation of chromatin fragments at the nuclear membrane, thereby impairing the activation of cGAS-STING and senescence-associated inflammation. Notably, nuclear egress of CCFs is inhibited by glucose limitation or metformin treatment. This is due to AMPK phosphorylation and autophagic degradation of the ESCRT-III component, ALIX. Metformin treatment in naturally aged mice downregulates ALIX protein and blocks cGAS activation and chronic inflammation in the small intestine. Together, our study defines a central mechanism linking nutrient sensing and chronic inflammation, two distinct hallmarks of aging, and suggests a new approach to suppress age-associated inflammation by targeting the nuclear egress of chromatin fragments.

cell biology↗

bioRxiv2024

Hippocampal CA2 neurons disproportionately express AAV-delivered genetic cargo

Hippocampal area CA2 is unique in many ways, largely based on the complement of genes expressed there. We and others have observed that CA2 neurons exhibit a uniquely robust tropism for adeno-associated viruses (AAVs) of multiple serotypes and variants. In this study, we aimed to systematically investigate the propensity for AAV tropism toward CA2 across a wide range of AAV serotypes and variants, injected either intrahippocampally or systemically, including AAV1, 2, 5, 6, 8, 9, DJ, PHP.B, PHP.eB, and CAP-B10. We found that most serotypes and variants produced disproportionally high expression of AAV-delivered genetic material in hippocampal area CA2, although two serotypes (AAV6 and DJ) did not. In an effort to understand the mechanism(s) behind this observation, we considered perineuronal nets (PNNs) that ensheathe CA2 pyramidal cells and, among other functions, buffer diffusion of ions and molecules. We hypothesized that PNNs might attract AAV particles and maintain them in close proximity to CA2 neurons, thereby increasing exposure to AAV particles. However, genetic deletion of PNNs from CA2 had no effect on AAV transduction. Next, we next considered the AAV binding factors and receptors known to contribute to AAV transduction. We found that the AAV receptor (AAVR), which is critical to transduction, is abundantly expressed in CA2, and knockout of AAVR nearly abolished expression of AAV-delivered material by all serotypes tested. Additionally, we found CA2 enrichment of several cell-surface glycan receptors that AAV particles attach to before interacting with AAVR, including heparan sulfate proteoglycans, N-linked sialic acid and N-linked galactose. Indeed, CA2 showed the highest expression of AAVR and the investigated glycan receptors within the hippocampus. We conclude that CA2 neurons are endowed with multiple factors that make it highly susceptible to AAV transduction, particularly to the systemically available PHP variants, including CAP-B10. Given the curved structure of hippocampus and the relatively small size of CA2, systemic delivery of engineered PHP or CAP variants could all but eliminate the need for intrahippocampal AAV injections, particularly when injecting recombinase-dependent AAVs into animals that express recombinases in CA2.

neuroscience↗

bioRxiv2024

Evaluation of T Cell Receptor Construction Methods from scRNA-Seq Data

T cell receptors (TCRs) serve pivotal roles in the adaptive immune system by enabling recognition and response to pathogens and irregular cells. Various methods exist for TCR construction from single-cell RNA sequencing (scRNA-seq) datasets, each with its unique characteristics regarding accuracy, sensitivity, adaptability, usability, time, and memory consumption. Yet, a comprehensive understanding of their relative strengths and weaknesses for different applications remains elusive. In our research, we implemented a benchmark analysis utilizing experimental single-cell immune profiling datasets encompassing paired scRNA-seq as input and scTCR-seq datasets as ground truth reference from human and mouse. Additionally, we introduced a novel simulator, YASIM-scTCR (Yet Another Simulator for single-cell TCR), capable of generating scTCR-seq reads containing a diverse array of TCR- derived sequences under different sequencing depths and read lengths. Our results consistently showed that TRUST4 outperformed others across multiple datasets, while MiXCR and DeRR also demonstrated considerable accuracy. We also discovered that the sequencing depth inherently imposes a critical constraint on successful TCR construction from scRNA-seq data. In summary, we present a benchmark study to aid researchers in choosing the most appropriate methods for reconstructing TCR from scRNA-seq data.

bioinformatics↗

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

Deep Learning Driven Cell-Type-Specific Embedding for Inference of Single-Cell Co-expression Networks

The inference of gene co-expression module in specific cell types is important for understanding of cell-type-specific biological processes. We have developed DeepCSCN, an unsupervised deep-learning framework, to infer gene co-expression modules from single-cell RNA sequencing (scRNA-seq) data. Utilizing a global-to-local network construction approach, DeepCSCN can infer co-expression at the whole sample level and construct cell-type-specific co-expression networks. Systematic evaluations on eight public scRNA-seq datasets show that DeepCSCN significantly outperforms eight existing methods in co-expression network construction. Furthermore, DeepCSCN effectively identifies cell-type-specific co-expression networks that are more enriched for cell-specific functional pathways compared to current methods. Finally, application of DeepCSCN on public scRNA-seq data revealed 280 cell-specific gene modules across 27 cell types, including epithelial cells, immune cells, and myonuclei, demonstrating its versatility and accuracy in elucidating cell-type-specific gene co-expression regulation. DeepCSCN offers a powerful tool for researchers to dissect the intricate gene co-expression networks within distinct cell types.

bioinformatics↗

bioRxiv2024

Transcranial focused ultrasound modulates visual thalamus in a nonhuman primate model

The thalamus plays a pivotal role as a neural hub, integrating and distributing visual information to cortical regions responsible for visual processing. Transcranial focused ultrasound (tFUS) has emerged as a promising non-invasive brain stimulation technology, enabling modulation of neural circuits with high spatial precision. This study investigates the tFUS neuromodulation at visual thalamus and characterizes the resultant effects on interconnected visual areas in a nonhuman primate model. Experiments were conducted on a rhesus macaque trained in a visual fixation task, combining tFUS stimulation with simultaneous scalp electroencephalography (EEG) and intracranial recordings from area V4, a region closely linked to the thalamus. Ultrasound was delivered through a 128-element random array ultrasound transducer operating at 700 kHz, with the focus steered onto the pulvinar of the thalamus based on neuroanatomical atlas and individual brain model. EEG source imaging revealed localized tFUS-induced activities in the thalamus, midbrain, and visual cortical regions. Critically, tFUS stimulation of the pulvinar can elicit robust neural responses in V4 without visual input, manifested as significant modulations in local field potentials, elevated alpha and gamma power, corroborating the functional thalamocortical connectivity. Furthermore, the tFUS neuromodulatory effects on visually-evoked V4 activities were region-specific within the thalamus and dependent on ultrasound pulse repetition frequency. This work provides direct electrophysiological evidence demonstrating the capability of tFUS in modulating the visual thalamus and its functional impact on interconnected cortical regions in a large mammalian model, paving the way for potential investigations for tFUS treating visual, sensory, and cognitive impairments.

neuroscience↗

bioRxiv2024

A Viral-Encoded Homologue of IPS-1 Modulates Innate Immune Signaling During KSHV Lytic Replication

Modulation of innate immunity is critical for virus persistence in a host. In particular, viral-encoded disruption of type I interferon, a major antiviral cytokine induced to fight viral infection, is a key component in the repertoire of viral pathogenicity genes. We have identified a previously undescribed open reading frame within the Kaposis sarcoma-associated herpesvirus (KSHV) genome that encodes a homologue of the human IPS-1 (also referred to as MAVS) protein that we have termed viral-IPS-1 (v-IPS-1). This protein is expressed during the lytic replication program of KSHV, and expression of v-IPS-1 blocks induction of type I interferon upstream of the TRAF3 signaling node including signaling initiated via both the RLR and TLR3/4 signaling axes. This disruption of signaling coincides with destabilization of the cellular innate signaling adaptors IPS-1 and TRIF along with a concatenate stabilization of the TRAF3 protein. Additionally, expression of v-IPS-1 leads to decreased antiviral responses indicating a blot to type I interferon induction during viral infection. Taken together, v-IPS-1 is the first described viral homologue of IPS-1 and this viral protein leads to reprogramming of innate immunity through modulation of type I interferon signaling during KSHV lytic replication.

microbiology↗

bioRxiv2024

Critical roles of Ikaros and HDAC1 in regulation of heterochromatin and tumor suppression in T-cell acute lymphoblastic leukemia

The IKZF1 gene encodes IKAROS - a DNA binding protein that acts as a tumor suppressor in T-cell acute lymphoblastic leukemia (T-ALL). IKAROS can act as a transcriptional repressor via recruitment of histone deacetylase 1 (HDAC1) and chromatin remodeling, however the mechanisms through which Ikaros exerts its tumor suppressor function via heterochromatin in T-ALL are largely unknown. We studied human and mouse T-ALL using a loss-of-function and IKZF1 re-expression approach, along with primary human T-ALL, and normal human and mouse thymocytes to establish the role of Ikaros and HDAC1 in global regulation of facultative heterochromatin and transcriptional repression in T-ALL. Results identified novel Ikaros and HDAC1 functions in T-ALL: Both Ikaros and HDAC1 are essential for EZH2 histone methyltransferase activity and formation of facultative heterochromatin; recruitment of HDAC1 by Ikaros is critical for establishment of H3K27me3 histone modification and repression of active enhancers; and Ikaros-HDAC1 complexes promote formation and expansion of H3K27me3 Large Organized Chromatin lysine (K) domains (LOCKs) and Broad Genic Repression Domains (BGRDs) in T-ALL. Our results establish the central role of Ikaros and HDAC1 in activation of EZH2, global regulation of the facultative heterochromatin landscape, and silencing of active enhancers that regulate oncogene expression.

bioinformatics↗

bioRxiv2024

Transcranial focused ultrasound activates feedforward and feedback cortico-thalamo-cortical pathways by selectively activating excitatory neurons

Transcranial focused ultrasound stimulation (tFUS) has been proven capable of altering focal neuronal activities and neural circuits non-invasively in both animals and humans. The abilities of tFUS for cell-type selection within the targeted area like somatosensory cortex have been shown to be parameter related. However, how neuronal subpopulations across neural pathways are affected, for example how tFUS affected neuronal connections between brain areas remains unclear. In this study, multi-site intracranial recordings were used to quantify the neuronal responses to tFUS stimulation at somatosensory cortex (S1), motor cortex (M1) and posterior medial thalamic nucleus (POm) of cortico-thalamo-cortical (CTC) pathway. We found that when targeting at S1 or POm, only regular spiking units (RSUs, putative excitatory neurons) responded to specific tFUS parameters (duty cycle: 6%-60% and pulse repetition frequency: 1500 and 3000 Hz ) during sonication. RSUs from the directly connected area (POm or S1) showed a synchronized response, which changed the directional correlation between RSUs from POm and S1. The tFUS induced excitation of RSUs activated the feedforward and feedback loops between cortex and thalamus, eliciting delayed neuronal responses of RSUs and delayed activities of fast spiking units (FSUs) by affecting local network. Our findings indicated that tFUS can modulate the CTC pathway through both feedforward and feedback loops, which could influence larger cortical areas including motor cortex.

neuroscience↗

bioRxiv2024

CD-GPT: A Biological Foundation Model Bridging the Gap between Molecular Sequences Through Central Dogma

The central dogma serves as a fundamental framework for understanding the flow and expression of genetic information within living organisms, facilitating the connection of diverse biological sequences across molecule types. In this study, we present CD-GPT (Central Dogma Generative Pretrained Transformer), a generative biological foundation model with 1 billion parameters, aiming to capture the sequence relationships between DNA, RNA, and proteins. We model sequences in a unified representational space and employ a shared, multi-molecule vocabulary to narrow their distances in the embedding space effectively. Through extensive pretraining on nucleotide and amino acid sequence data, CD-GPT exhibits exceptional performance in a wide range of predictive and generative downstream tasks, including mono-molecular and multi-molecular analyses. Notably, CD-GPT excels in tasks such as genomic element detection, protein property prediction, RNA-protein interaction identification and also generative tasks like protein generation and reverse translation. The versatility of CD-GPT opens up promising avenues for advanced multi-omics analysis.

bioinformatics↗

bioRxiv2024

An Investigation of Parameter-Dependent Cell-Type Specific Effects of Transcranial Focused Ultrasound Stimulation Using an Awake Head-Fixed Rodent Model

Transcranial focused ultrasound (tFUS) is a promising neuromodulation technique able to target shallow and deep brain structures with high precision. Previous studies have demonstrated that tFUS stimulation responses are both cell-type specific and controllable through altering stimulation parameters. Specifically, tFUS can elicit time-locked neural activity in regular spiking units (RSUs) that is sensitive to increases in pulse repetition frequency (PRF), while time-locked responses are not seen in fast spiking units (FSUs). These findings suggest a unique capability of tFUS to alter circuit network dynamics with cell-type specificity; however, these results could be biased by the use of anesthesia, which significantly modulates neural activities. In this study, we develop an awake head-fixed rat model specifically designed for tFUS study, and address a key question if tFUS still has cell-type specificity under awake conditions. Using this novel animal model, we examined a series of PRFs and burst duty cycles (DCs) to determine their effects on neuronal subpopulations without anesthesia. We conclude that cell-type specific time-locked and delayed responses to tFUS as well as PRF and DC sensitivity are present in the awake animal model and that despite some differences in response, isoflurane anesthesia is not a major confound in studying the cell-type specificity of ultrasound neuromodulation. We further determine that, in an awake, head-fixed setting, the preferred PRF and DC for inducing time-locked excitation with our pulsed tFUS paradigm are 1500 Hz and 60%, respectively.

neuroscience↗

bioRxiv2024

Comparative genomics of three non-hematophagous leeches (Whitmania spp.): focusing on antithrombotic genes

Leeches are well known for their blood-feeding habits and are widely used for medicinal purposes as they secrete various antithrombotic substances. However, some leeches such as Whitmania spp. exhibit non-hematophagous feeding habits and their significance for medicinal use is debated. In this study, we provide chromosome-level genomes of two non-hematophagous leeches Whitmania acranulata and Whitmania laevis, and combined with our previous results of Whitmania pigra, we systematically analyzed the similarities and differences on the genomes and especially their antithrombotic genes among the three non-hematophagous Whitmania leeches. For W. acranulata, W. laevis, and W. pigra, the genome size (181.72 Mb, 173.87 Mb, and 173.56), the percentage of repeat sites (29.55%, 28.28%, and 27.02%), and the number of protein-coding genes (27,068, 23,805, and 24,156) were close to each other, respectively. In contrast, both the total number of the antithrombotic genes (100, 63, and 79), and the detailed constitutes of different antithrombotic gene families were obviously different among the three leeches. There were also massive genetic variations among the members within each antithrombotic gene/protein family. RNA-Seq-based gene expression estimation showed that the expression profiles of the antithrombotic gene families were apparently different among the three leeches. This is the most comprehensive comparison of the genomes and antithrombic biomacromolecules for the Whitmania leeches to date. Our results will greatly facilitate the evolutionary research and application of leech derivatives for medical and pharmaceutical purposes of thrombosis.

genomics↗

bioRxiv2024

Spontaneous HFO Sequences Reveal Propagation Pathways for Precise Delineation of Epileptogenic Networks

AbstractEpilepsy, a neurological disorder affecting millions worldwide, poses great challenges in precisely delineating the epileptogenic zone - the brain region generating seizures - for effective treatment. High-frequency oscillations (HFOs) are emerging as promising biomarkers; however, the clinical utility is hindered by the difficulties in distinguishing pathological HFOs from non- epileptiform activities at single electrode and single patient resolution and understanding their dynamic role in epileptic networks. Here, we introduce an HFO-sequencing approach to analyze spontaneous HFOs traversing cortical regions in 40 drug-resistant epilepsy patients. This data- driven method automatically detected over 8.9 million HFOs, pinpointing pathological HFO- networks, and unveiled intricate millisecond-scale spatiotemporal dynamics, stability, and functional connectivity of HFOs in prolonged intracranial EEG recordings. These HFO sequences demonstrated a significant improvement in localization of epileptic tissue, with an 818.47% increase in concordance with seizure-onset zone (mean error: 2.92 mm), compared to conventional benchmarks. They also accurately predicted seizure outcomes for 90% AUC based on pre-surgical information using generalized linear models. Importantly, this mapping remained reliable even with short recordings (mean standard deviation: 3.23 mm for 30-minute segments). Furthermore, HFO sequences exhibited distinct yet highly repetitive spatiotemporal patterns, characterized by pronounced synchrony and predominant inward information flow from periphery towards areas involved in propagation, suggesting a crucial role for excitation-inhibition balance in HFO initiation and progression. Together, these findings shed light on the intricate organization of epileptic network and highlight the potential of HFO-sequencing as a translational tool for improved diagnosis, surgical targeting, and ultimately, better outcomes for vulnerable patients with drug-resistant epilepsy. One Sentence SummaryPathological fast brain oscillations travel like traffic along varied routes, outlining recurrently visited neural sites emerging as critical hotspots in epilepsy network.

neuroscience↗

bioRxiv2024

Biologically informed machine learning modeling of immune cells to reveal physiological and pathological aging process

The immune system undergoes progressive functional remodeling from neonatal stages to old age. Therefore, understanding how aging shapes immune cell function is vital for precise treatment of patients at different life stages. Here, we constructed the first transcriptomic atlas of immune cells encompassing human lifespan, ranging from newborns to supercentenarians, and comprehensively examined gene expression signatures involving cell signaling, metabolism, differentiation, and functions in all cell types to investigate immune aging changes. By comparing immune cell composition among different age groups, HLA highly expressing NK cells and CD83 positive B cells were identified with high percentages exclusively in the teenager (Tg) group, whereas CD4_CTL precursors were exclusively enriched in the supercentenarian (Sc) group. Notably, we found that the biological age (BA) of pediatric COVID-19 patients with multisystem inflammatory syndrome accelerated aging according to their chronological age (CA). Besides, we proved that inflammatory shift-myeloid abundance and signature correlate with the progression of complications in Kawasaki disease (KD). Finally, based on those age-related immune cell compositions, we developed a novel BA prediction model, PHARE (https://xiazlab.org/phare/), which applies to both scRNA-seq and bulk RNA-seq data. Overall, our study revealed changes in immune cell proportions and function associated with aging, both in health and disease, and provided a novel tool for successfully capturing features that accelerate or delay aging.

immunology↗

bioRxiv2024

Structural insights into physiological activation and antagonism of melanin-concentrating hormone receptor MCHR1

Melanin-concentrating hormone (MCH) is a 19-amino-acid neuropeptide playing crucial roles in energy homeostasis, sleep, and various physiological processes. It acts through two G protein-coupled receptors, MCHR1 and MCHR2, with MCHR1 being universally present in mammals and a potential target for treating metabolic and mental health conditions. However, drug development efforts have been impeded by the lack of structural information. Here, we present the cryo-EM structures of MCHR1 in its active state complexed with MCH and Gi1, as well as in its inactive state bound to a selective antagonist SNAP-94847. Structural and mutagenesis analyses disclosed the recognition mechanisms for both MCH and SNAP-94847, the activation mechanism and antagonism of MCHR1, and the determinants of ligand specificity. These findings are expected to accelerate the development of better drugs targeting the MCH system.

biophysics↗

bioRxiv2024

Fast and accurate modeling and design of antibody-antigen complex using tFold

Accurate prediction of antibody-antigen complex structures holds significant potential for advancing biomedical research and the design of therapeutic antibodies. Currently, structure prediction for protein monomers has achieved considerable success, and promising progress has been made in extending this achievement to the prediction of protein complexes. However, despite these advancements, fast and accurate prediction of antibody-antigen complex structures remains a challenging and unresolved issue. Existing end-to-end prediction methods, which rely on homology and templates, exhibit sub-optimal accuracy due to the absence of co-evolutionary constraints. Meanwhile, conventional docking-based methods face difficulties in identifying the contact interface between the antigen and antibody and require known structures of individual components as inputs. In this study, we present a fully end-to-end approach for three-dimensional (3D) atomic-level structure predictions of antibodies and antibody-antigen complexes, referred to as tFold-Ab and tFold-Ag, respectively. tFold leverages a large protein language model to extract both intra-chain and inter-chain residue-residue contact information, as well as evolutionary relationships, avoiding the time-consuming multiple sequence alignment (MSA) search. Combined with specially designed modules such as the AI-driven flexible docking module, it achieves superior performance and significantly enhanced speed in predicting both antibody (1.6% RMSD reduction in the CDR-H3 region, thousand times faster) and antibody-antigen complex structures (37% increase in DockQ score, over 10 times faster), compared to AlphaFold-Multimer. Given the performance and speed advantages, we further extend the capability of tFold for structure-based virtual screening of binding antibodies, as well as de novo co-design of both structure and sequence for therapeutic antibodies. The experiment results demonstrate the potential of tFold as a high-throughput tool to enhance processes involved in these tasks. To facilitate public access, we release code and offer a web service for antibody and antigen-antibody complex structure prediction, which is available at https://drug.ai.tencent.com/en.

bioinformatics↗