Biology subjects

Yang, X.

Publications and source records attributed to Yang, X..

At least 451 records · Page 25Linked to original sources

bioRxiv2023

Three-dimensional visualization of blood vessels in human gliomas based on tissue clearing and deep learning

Gliomas, with their intricate and aggressive nature, call for a detailed visualization of their vasculature. While many studies lean towards 2D imaging of thin sections, this method often overlooks the full spatial heterogeneity inherent to tumors. To overcome this limitation, our study melded state-of-the-art techniques, encompassing tissue clearing technology, 3D confocal microscopy imaging, and deep learning-aided vessel extraction, resulting in a comprehensive 3D visualization of glioma vasculature in intact human tissue. Specifically, we treated formalin-fixed thick human glioma tissue sections (500 m) with OPTIClear for transparency and subsequently performed immunofluorescent labeling using CD31. Using confocal microscopy, we obtained 3D images of the glioma vasculature. For vessel extraction, we employed a specialized 3D U-Net, enriched with image preprocessing and post-processing methods, and benchmarked its performance against the Imaris software. Our findings indicated that OPTIClear-enabled tissue clearing yielded a holistic 3D representation of immunolabeled vessels in clinical human glioma samples. Impressively, our deep learning technique outshined the traditional Imaris approach in terms of accuracy and efficiency in vessel extraction. Further, discernible variations in vascular metrics, such as mean diameter, branching point count, and volume ratio, were observed between low-grade and high-grade gliomas. In essence, our innovative blend of tissue clearing and deep learning not only enables enhanced 3D visualization of human glioma vasculature but also underscores morphological disparities across glioma grades, potentially influencing pathological grading, therapeutic strategies, and prognostic evaluations.

neuroscience↗

bioRxiv2023

Visual boundary cues, alone, can anchor place and grid cells in virtual reality

The hippocampal formation contains neurons responsive to an animals current location and orientation, which together provide the organism with a neural map of space1-3. Spatially tuned neurons rely on external landmark cues and internally generated movement information to estimate position4,5. An important class of landmark cue are the boundaries delimiting an environment, which can define place cell field position6,7 and stabilise grid cell firing8. However, the precise nature of the sensory information used by the hippocampus to detect boundaries remains unknown. Here we use 2-dimensional virtual reality (VR)9 to show that purely visual cues from the surrounding walls are both sufficient and necessary to stabilise place and grid cell responses. Unstable grid cells retain internally coherent, hexagonally arranged firing fields, but these fields drift with respect to the virtual environment over periods >5 sec. Optic flow from a virtual floor does not slow drift dynamics, emphasising the importance of boundary-related visual information. Surprisingly, place fields are more stable close to boundaries even when these are invisible. Invisible boundary locations can only be inferred from the relative movements of the animal and a reward beacon, specifically, from the mismatch between observed beacon movement and that predicted from motor output. This shows that boundary information can be detected through sensorimotor feedback. We find that subsets of place cells show allocentric directional tuning towards the beacon, with strength of tuning correlating with place field stability when boundaries are removed.

neuroscience↗

bioRxiv2023

Cell-type-resolved somatic mosaicism reveals clonal dynamics of the human forebrain

Debate remains around anatomic origins of specific brain cell subtypes and lineage relationships within the human forebrain. Thus, direct observation in the mature human brain is critical for a complete understanding of the structural organization and cellular origins. Here, we utilize brain mosaic variation within specific cell types as distinct indicators for clonal dynamics, denoted as cell-type-specific Mosaic Variant Barcode Analysis. From four hemispheres from two different human neurotypical donors, we identified 287 and 780 mosaic variants (MVs), respectively that were used to deconvolve clonal dynamics. Clonal spread and allelic fractions within the brain reveal that local hippocampal excitatory neurons are more lineage-restricted compared with resident neocortical excitatory neurons or resident basal ganglia GABAergic inhibitory neurons. Furthermore, simultaneous genome-transcriptome analysis at both a cell-type-specific and single-cell level suggests a dorsal neocortical origin for a subgroup of DLX1+ inhibitory neurons that disperse radially from an origin shared with excitatory neurons. Finally, the distribution of MVs across 17 locations within one parietal lobe reveals restrictions of clonal spread in the anterior-posterior axis precedes that of the dorsal-ventral axis for both excitatory and inhibitory neurons. Thus cell-type resolved somatic mosaicism can uncover lineage relationships governing the development of the human forebrain.

neuroscience↗

bioRxiv2023

Molecular mechanism of α-synuclein aggregation on lipid membranes revealed.

The central hallmark of Parkinsons disease pathology is the aggregation of the -synuclein protein, which, in its healthy form, is associated with lipid membranes. Purified monomeric -synuclein is relatively stable in vitro, but its aggregation can be triggered by the presence of lipid vesicles. Despite this central importance of lipids in the context of -synuclein aggregation, their mechanistic role in this process has not been established to date. Here, we use chemical kinetics to develop a detailed mechanistic model that is able to globally describe the aggregation behaviour of -synuclein in the presence of DMPS lipid vesicles, across a range of lipid and protein concentrations. Through the application of our kinetic model to experimental data, we find that the reaction is a co-aggregation process involving both protein and lipids and that lipids promote aggregation predominantly by enabling the elongation process. Moreover, we find that the initial formation of aggregates, via primary nucleation, takes place not on the surface of lipid vesicles but at the interfaces present in vitro. Our model will enable mechanistic insights, also in other lipid-protein co-aggregation systems, which will be crucial in the rational design of drugs that inhibit aggregate formation and act at the key points in the -synuclein aggregation cascade.

biophysics↗

bioRxiv2023

Architecture and activation of human muscle phosphorylase kinase

The study of phosphorylase kinase (PhK)-regulated glycogen metabolism has contributed to the fundamental understanding of protein phosphorylation. Here we present the high-resolution cryo-electron microscopy structures of human muscle PhK. The 1.3-megadalton PhK 4{beta}4{gamma}4{delta}4 hexadecamer consists of a tetramer of tetramer, wherein four {beta}{gamma}{delta} modules are connected by the central {beta}4 scaffold. The - and {beta}-subunits possess glucoamylase-like domains, but exhibit no detectable enzyme activities. The -subunit serves as a bridge between the {beta}-subunit and the {gamma}{delta} subcomplex, and facilitates the {gamma}-subunit to adopt an autoinhibited state. Ca2+-free calmodulin ({delta}-subunit) binds to the {gamma}-subunit in a compact conformation. Upon binding of Ca2+, a conformational change occurs, allowing for the de-inhibition of the {gamma}-subunit through a spring-loaded mechanism. We also reveal an ADP-binding pocket in the {beta}-subunit, which plays a role in allosterically enhancing PhK activity. These results provide unprecedented molecular insights of this important kinase complex.

biochemistry↗

bioRxiv2023

Flap endonuclease 1 repairs DNA-protein crosslinks via ADP-ribosylation

DNA-protein crosslinks (DPCs) are among the most ubiquitous and detrimental DNA lesions which arise from exposure to metabolic stresses, drugs, or crosslinking agents such as formaldehyde (FA). FA is a cellular by-product of methanol metabolism, histone demethylation, lipid peroxidation as well as environmental pollutants. Failure to repair FA-induced DPCs blocks nearly all chromatin-based processes including replication and transcription, leading to immunodeficiencies, neurodegeneration, and cancer. Yet, it remains largely unknown how the cell repairs DPCs. The study of DPC repair is impeded by our incomprehension of the types of proteins crosslinked by FA due to the lack of techniques to identify the DPCs. Here, we designed a novel bioassay to profile FA-induced DPCs by coupling cesium chloride differential ultracentrifugation with HPLC-mass spectrometry (MS). Using the method, we revealed the proteome of FA-induced DPCs in human cells and found that the most abundant proteins that form DPCs are PARP1, topoisomerases I and II, methyltransferases, DNA and RNA polymerases, histones, as well as ribosomal proteins. To identify enzymes that repair DPCs, we carried out RNA interference screening and found that downregulation of flap endonuclease 1 (FEN1) rendered cells hypersensitive to FA. Since FEN1 possesses 5-flap endonuclease activity, we hypothesized that FA induces DPC-conjugated 5-flap DNA fragments that can be processed by FEN1. Indeed, we demonstrate that FA damages DNA bases that are converted into 5-flap via the base excision pathway (BER). We also observed that the damaged DNA bases were colocalized with DPCs and FEN1. Mechanistically, we showed that FEN1 repairs FA-induced DPCs in vivo and cleaves 5-flap DNA substrate harboring DPC mimetic in vitro. We also found that FEN1 repairs enzymatic topoisomerase II (TOP2)-DPCs induced by their inhibitors etoposide and doxorubicin independently of the BER pathway, and that FEN1 and the DPC-targeting protease SPRTN act as parallel pathways for the repair of both FA-induced non-enzymatic DPCs and etoposide-induced enzymatic TOP2-DPCs. Notably, we found that FA-induced non-enzymatic DPCs and enzymatic TOP2-DPCs are promptly modified by poly-ADP-ribosylation (PARylation), a post-translational modification catalyzed by PARP1, a key DNA damage response effector that acts by PARylating both DNA damage sites and DNA repair proteins. We performed immunoprecipitation (IP) assays with anti-PAR antibody for HPLC-MS and identified FEN1 as a PARylation substrate. Next, we showed that PARylation of DPC substrates signaled FEN1 whereas PARylation of FEN1 drove FEN1 to DPC sites. Finally, using the enzymatic labeling of the terminal ADP-ribose-MS method, we identified the E285 residue of FEN1 as a dominant PARylation site, which appeared to be required for FEN1 relocation to DPCs. Taken together, our work not only unveiled the identities of FA-induced DPCs but also discovered an unprecedented PARP1-FEN1 nuclease pathway as a universal and imperative mechanism to repair the miscellaneous DPCs and prevent DPC-induced genomic instability.

molecular biology↗

bioRxiv2023

Pan-cancer evolution signatures link clonal expansion to dynamic changes in the tumour immune microenvironment

Cancer is an evolutionary process characterised by profound intra-tumour heterogeneity. Intra-tumour heterogeneity can be quantified using in silico estimates of cancer cell fractions of tumour-specific somatic mutations. Here we demonstrate a data-driven approach that uses cancer cell fraction distributions to identify 4 robust pan-cancer evolutionary signatures from an analysis of 4,146 individual tumour samples (TCGA) representing 17 distinct cancer types. Evolutionary signatures defined a continuum of cancer cell fractions representing neutral evolution, clonal expansion and fixation. Correlation of evolutionary signatures with programs representing distinct mutational and biological processes demonstrated that individual tumours enriched for clonal expansions and fixations were associated with immune evasion and distinct changes in the tumour immune microenvironment. We observed a dynamic switch between adaptive and innate immune processes as tumours undergo clonal fixation and escape immune surveillance. We also identify mutational processes underpinning different modes of tumour evolution and demonstrate that switching between adaptive and innate immune cell populations is accompanied by the clonal expansion of driver genes that modulate tumour-stroma interactions1.

bioinformatics↗

bioRxiv2023

Engineered dsRNA-protein nanoparticles for effective long-distance transport, delivery and gene silencing in plants

Long-distance transport of exogenous biologically active RNA molecules in higher plants has not been reported. Here, we report that cationized bovine serum albumin (cBSA) avidly binds double-stranded beta-glucuronidase RNA (dsGUS RNA) to form nucleic acid-protein nanocomplexes. Using tobacco and poplar plants, we have shown effective uptake and long-distance transport of cBSA/dsGUS RNA nanocomplexes from basal ends of leaf petioles to leaf blades as well as from basal ends of shoots to their apexes and apical leaves. We have further demonstrated that the cBSA/dsGUS RNA nanocomplexes are highly effective in silencing both conditionally inducible DR5-GUS gene and constitutively active 35S-GUS gene in leaf, shoot and shoot meristem tissues. This cBSA/dsRNA delivery technology may provide a convenient, fast, and inexpensive tool for characterizing gene functions in plants, and potentially for in planta gene-editing.

molecular biology↗

bioRxiv2023

MSIsensor-RNA: microsatellite instability detection for bulk and single-cell gene expression data

Microsatellite instability (MSI) is an indispensable biomarker in cancer immunotherapy. Currently, MSI scoring methods by high-throughput omics methods have gained popularity and demonstrated better performance than the gold standard method for MSI detection. However, the MSI detection method on expression data, especially single-cell expression data, is still lacking, limiting the scope of clinical application and prohibiting the investigation of MSI at a single-cell level. Herein, we developed MSIsensor-RNA, an accurate, robust, adaptable, and standalone software to detect MSI status based on expression values of MSI-associated genes. We demonstrated the favorable performance and promise of MSIsensor-RNA in both bulk and single-cell gene expression data in multiplatform technologies including RNA-seq, microarray, and single-cell RNA-seq. MSIsensor-RNA is a versatile, efficient, and robust method for MSI status detection from both bulk and single-cell gene expression data in clinical studies and applications. MSIsensor-RNA is available at https://github.com/xjtu-omics/msisensor-rna.

bioinformatics↗

bioRxiv2023

Pindel-TD: a tandem duplication detector based on a pattern growth approach

Tandem duplication (TD) is a major type of structural variation (SV), and plays an important role in novel gene formation and human diseases. However, TDs are often missed or incorrectly classified as insertions by most of modern SV detection methods due to the lacking of specialized operation on TD related mutational signals. Herein, we developed a TD detection module of Pindel referred as Pindel-TD based on a TD specific pattern growth approach. Pindel-TD detects TDs with a wide size range at single nucleotide resolution. Using simulation and real read data of HG002, we demonstrate that Pindel-TD outperformed other leading methods in terms of precision, recall, F1-score and robustness. Further applying Pindel-TD on data generated from K562 cancer cell line, we identified a TD located at the seventh exon of SAGE1, explaining its high expression. Pindel-TD is available at https://github.com/xjtu-omics/pindel and free for non-commercial use.

bioinformatics↗

bioRxiv2023

Tim-1 and Tim-4 mediate entry of the human Kaposi's sarcoma-associated herpesvirus and the related rhesus monkey rhadinovirus

Kaposis sarcoma-associated herpesvirus (KSHV) is a human tumor virus. It is associated with Kaposis sarcoma, primary effusion lymphoma, and multicentric Castlemans disease. KSHV is known to interact with several different receptors, among them heparan sulfate proteoglycans, Eph family receptors, and integrins. We mutated the closely related rhesus monkey rhadinovirus in the known receptor interaction sites for Eph family and Plexin domain containing proteins and found it to still replicate on certain cells. This lytic virus was then used as a selection agent in a genome-wide CRISPR knockout screen, which identified TIM1 and NRP1 as host dependency factors. NRP1 is also host factor for the related Epstein-Barr virus and was recently reported to promote KSHV infection, which we confirm even if it functions with low efficiency on most cells and became functional only after ablation of the Eph receptor interaction. Further analysis through overexpression demonstrated that Tim-1 and the related Tim-4 are strong mediators of RRV and KSHV infection, in particular in the absence of other receptor interactions and even more pronounced for a KSHV mutant deleted in glycoprotein K8.1. Both Tim-1 and Tim-4 are heavily O-glycosylated phosphatidylserine (PS) receptors. For KSHV in particular, experiments with mutated Tim-1 and comparison to Ebola virus glycoprotein-driven entry indicate that the interaction with Tim-1 occurs through PS-binding by Tim-1 and suggest additional interaction in a PS-independent manner. The mucin-like domain of Tim-1 is required for optimal receptor function. The use of Tim proteins for entry is a novelty for herpesviruses and underscores the unique biology of KSHV and RRV.

microbiology↗

bioRxiv2023

Large-scale imputation models for multi-ancestry proteome-wide association analysis

Proteome-wide association studies (PWAS) decode the intricate proteomic landscape of biological mechanisms for complex diseases. Traditional PWAS model training relies heavily on individual-level reference proteomes, restricting its capacity to harness the emerging summary-level protein quantitative trait loci (pQTL) data in the public domain. Here we introduced BLISS, a novel framework to train protein imputation models using only pQTL summary statistics. By leveraging extensive pQTL data from the UK Biobank, deCODE, and ARIC studies, we applied BLISS to develop large-scale European PWAS models covering 5,779 unique proteins. We further extended BLISS to integrate with small-scale non-European individual-level datasets, enabling the development of models tailored to Asian and African ancestries. We validated the performance of BLISS models through a systematic multi-ancestry analysis of over 2,500 phenotypes across five major genetic data resources. The newly identified protein-phenotype associations offer valuable insights into their cross-ancestry transferability, the contributions of different proteomic platforms, and the complementary perspectives provided by distinct genomic mapping approaches relevant to drug discovery. The developed models and data resources are freely available at https://www.gcbhub.org/.

genetics↗

bioRxiv2023

Staphyloccocus aureus biofilm, in the absence of planktonic bacteria, produces factors that activate counterbalancing inflammatory and immune-suppressive genes in human monocytes

Staphyloccocus aureus (S. aureus) is a major bacterial pathogen in orthopedic periprosthetic joint infection (PJI). S. aureus forms biofilms that promote persistent infection by shielding bacteria from immune cells and inducing an antibiotic-resistant metabolic state. We developed an in vitro system to study S. aureus biofilm interactions with primary human monocytes in the absence of planktonic bacteria. In line with previous in vivo data, S. aureus biofilm induced expression of inflammatory genes such as TNF and IL1B, and their anti-inflammatory counter-regulator IL-10. S. aureus biofilm also activated expression of PD-1 ligands that suppress T cell function, and of IL-1RA that suppresses differentiation of protective Th17 cells. Gene induction did not require monocyte:biofilm contact and was mediated by a soluble factor(s) produced by biofilm-encased bacteria that was heat resistant and > 3 kD in size. Activation of suppressive genes by biofilm was sensitive to suppression by Jak inhibition. These results support an evolving paradigm that biofilm plays an active role in modulating immune responses, and suggest this occurs via production of a soluble vita-PAMP. Induction of T cell suppressive genes by S. aureus biofilm provides insights into mechanisms that suppress T cell immunity in PJI, and suggest that anti-PD-1 therapy that is modeled on immune checkpoint blockade for tumors may be beneficial in PJI.

immunology↗

bioRxiv2023

Single-cell impedance cytometry of anticancer drug-treated tumor cells exhibiting mitotic arrest state to apoptosis using low-cost silver-PDMS microelectrodes

Chemotherapeutic drugs such as paclitaxel and vinblastine interact with the microtubules, and thus induce complex cell states of mitosis arrest at the G2/M phase followed by apoptosis dependent on drug exposure time and concentration. Microfluidic impedance cytometry (MIC) as a label-free and high-throughput technology for single-cell analysis, has been applied for viability assay of cancer cells post drug exposure at fixed time and dosage, yet verification of this technique for varied tumor cell states after anticancer drug treatment remains vacant. Here we present a novel MIC device and for the first time perform impedance cytometry on carcinoma cells exhibiting progressive states of G2/M arrest followed by apoptosis related to drug concentration and exposure time, after treatments by paclitaxel and vinblastine, respectively. Our results from impedance cytometry reveal increased amplitude and negative phase shift at low frequency, as well as higher opacity for the Hela cells under G2/M mitotic arrest compared to the untreated cells. The cells under apoptosis, on the other hand, exhibit opposite changes in these electrical parameters. Therefore, the impedance features differentiate the Hela cells under progressive states post anticancer drug treatment. We also demonstrate that vinblastine poses a more potent drug effect than paclitaxel especially at low concentrations. Our device is fabricated with a unique sacrificial layer-free soft lithography process as compared to the existing MIC device, which gives rise to readily aligned parallel microelectrodes made of silver-PDMS embedded in PDMS channel sidewalls with one molding step. Our results uncover the potential of the MIC device, with a fairly simple and low-cost fabrication process, for cellular state screening in anticancer drug therapy.

cell biology↗

bioRxiv2023

Uncovering the Roles of Mycobacterium tuberculosis melH in Redox and Bioenergetic Homeostasis: Implications for Antitubercular Therapy

Mycobacterium tuberculosis (Mtb), the pathogenic bacterium that causes tuberculosis, has evolved sophisticated defense mechanisms to counteract the cytotoxicity of reactive oxygen species (ROS) generated within host macrophages during infection. The melH gene in Mtb and Mycobacterium marinum (Mm) plays a crucial role in defense mechanisms against ROS generated during infection. We demonstrate that melH encodes an epoxide hydrolase and contributes to ROS detoxification. Deletion of melH in Mm resulted in a mutant with increased sensitivity to oxidative stress, increased accumulation of aldehyde species, and decreased production of mycothiol and ergothioneine. This heightened vulnerability is attributed to the increased expression of whiB3, a universal stress sensor. The absence of melH also resulted in reduced intracellular levels of NAD+, NADH, and ATP. Bacterial growth was impaired, even in the absence of external stressors, and the impairment was carbon-source-dependent. Initial MelH substrate specificity studies demonstrate a preference for epoxides with a single aromatic substituent. Taken together, these results highlight the role of melH in mycobacterial bioenergetic metabolism and provide new insights into the complex interplay between redox homeostasis and generation of reactive aldehyde species in mycobacteria. ImportanceThis study unveils the pivotal role played by the melH gene in Mycobacterium tuberculosis and Mycobacterium marinum in combatting the detrimental impact of oxidative conditions during infection. This investigation revealed notable alterations in the level of cytokinin-associated aldehyde, para-hydroxybenzaldehyde, as well as the redox buffer ergothioneine, upon deletion of melH. Moreover, changes in crucial cofactors responsible for electron transfer highlighted melHs crucial function in maintaining a delicate equilibrium of redox and bioenergetic processes. MelH prefers epoxide small substrates with a phenyl substituted substrate. These findings collectively emphasize the potential of melH as an attractive target for the development of novel antitubercular therapies that sensitize mycobacteria to host stress, offering new avenues for combating tuberculosis.

microbiology↗

bioRxiv2023

Novel Epidermal Oxysterols Function as Alarm Substances in Zebrafish

Aquatic animals often use chemical cues to signal predation risk. When injured, shoaling fish skins release alarm substances that induce intense fear and a suite of anti-predator behaviors in other shoal members. However, the chemical nature of alarm substances remains largely unknown. Here we show that zebrafish alarm substance comprises 24-methyl-5-cholestane-3,7,12,24,28-pentahydroxy 28-sulfate, a novel oxysterol sulfate, and 5-cyprinol sulfate. These compounds are present in zebrafish skin extract and, at concentrations of less than one nanomolar, each induces anti-predator behaviors and increases cortisol levels. Their mixture, at its natural ratio, replicates the skin extract in eliciting the full suite of anti-predator behavior patterns. Our findings reveal a molecular-level mechanism whereby fish signal predation danger.

animal behavior and cognition↗

bioRxiv2023

Central pore-opening structure and gating of the mechanosensitive PIEZO1 channel

PIEZO1 is a mechanically activated cation channel that undergoes force-induced activation and inactivation to control diverse physiological processes1-4. However, the distinct functional states and gating transitions remain structurally undefined. In response to changes in membrane forces, PIEZO1 transitions from a curved to a flattened shape, which might lead to opening and subsequent inactivation of the ion-conducting pore5,6. Here we employed a PIEZO1 mutant to capture various conformational states including an intermediate state with an opened central pore, allowing us to reveal the gating dynamics of this multiple-gate system. Upon transition from the curved to the partially flattened intermediate state, the trimeric cap undergoes a downward and rotational motion to open the lateral cap-gate, while the loop that connects the cap to the pore-lining inner helix acts as a spring-linker, switching from an extended to a compressed state. Along with the partial flattening of the peripheral blades, these conformational changes collectively dilate the cap-gate, spring-gate and the hydrophobic transmembrane-gate, leading to ion permeation through the transmembrane pore. Upon inactivation, while the blades remain flattened, the cap and spring-linker return to an up and extended state. Molecular dynamics simulations verify the ion-permeating pathway and the distinct ion-conducting states of the assigned close, intermediate, open, and inactivated structures. Mutagenesis and electrophysiological characterizations reveal that domains and residues involved in conformational dynamics of the cap and spring-linker are critical for mechanical activation of PIEZO1. Together, these studies have structurally defined the distinct functional states and the curvature-based activation and inactivation mechanisms of PIEZO1.

neuroscience↗

bioRxiv2023

GeneCompass: Deciphering Universal Gene Regulatory Mechanisms with Knowledge-Informed Cross-Species Foundation Model

Deciphering the universal gene regulatory mechanisms in diverse organisms holds great potential to advance our knowledge of fundamental life process and facilitate research on clinical applications. However, the traditional research paradigm primarily focuses on individual model organisms, resulting in limited collection and integration of complex features on various cell types across species. Recent breakthroughs in single-cell sequencing and advancements in deep learning techniques present an unprecedented opportunity to tackle this challenge. In this study, we developed GeneCompass, the first knowledge-informed, cross-species foundation model pre-trained on an extensive dataset of over 120 million single-cell transcriptomes from human and mouse. During pre-training, GeneCompass effectively integrates four types of biological prior knowledge to enhance the understanding of gene regulatory mechanisms in a self-supervised manner. Fine-tuning towards multiple downstream tasks, GeneCompass outperforms competing state-of-the-art models in multiple tasks on single species and unlocks new realms of cross-species biological investigation. Overall, GeneCompass marks a milestone in advancing knowledge of universal gene regulatory mechanisms and accelerating the discovery of key cell fate regulators and candidate targets for drug development.

bioinformatics↗