Biology subjects

Liu, B.

Publications and source records attributed to Liu, B..

At least 163 records · Page 9Linked to original sources

bioRxiv2024

Genetic variation in the species Arabidopsis thaliana reveals the existence of natural heat resilience factors for meiosis

Heat interferes with multiple meiotic processes leading to genome instability and sterility in flowering plants, including many crops. Despite its importance for food security, the mechanisms underlying heat tolerance of meiosis are poorly understood. In this study, we analyzed different meiotic processes in the Arabidopsis (Arabidopsis thaliana) accessions Columbia (Col) and Landsberg erecta (Ler), their F1 hybrids and F2 offspring under heat stress (37{degrees}C). At 37{degrees}C, Col exhibits significantly reduced formation of double-stand breaks (DSBs) and completely abolished homolog pairing, synapsis and crossover (CO) formation. Strikingly, Ler and Ler/Col hybrids are much less affected than Col. Interestingly, only 10% [~] 20% of F2 offspring exhibit the same heat tolerance of meiotic recombination as parents, indicating that heat resilience in Ler is controlled by the interplay of several loci. Moreover, F2 offspring show defective chromosome condensation in interkinesis, and untimely sister-chromatid segregation and/or chromosome fragmentation, the levels of which exceed those in either inbreds and/or hybrids thus implying a transgressive effect on heat tolerance of meiosis. Furthermore, correlation and cytogenetic analysis suggest that homolog pairing and/or synapsis have an impact on heat tolerance of chromosome morphology and stability during post-recombination stages under heat stress. Taken together, this study reveals the existence of natural heat resilience factors for meiosis in Arabidopsis, which have the great potential to be exploited in breeding programs. Author summaryEnvironmental temperature alterations affect meiotic recombination and/or chromosome segregation thus perturbing genetic makeup and genome stability in plants. We have previously reported that CO formation is fully abolished in Arabidopsis thaliana accession Col under heat stress (36{degrees}C-38{degrees}C) due to reduced DSB formation and impaired homolog pairing. Here, we show that in Arabidopsis thaliana accession Ler under the same high temperature conditions, both DSB and CO formation occur normally, and homolog pairing is mildly impacted, which indicate a striking difference in heat tolerance of meiotic recombination from Col. Remarkably, Col/Ler hybrids display the same heat tolerance as Ler, however, only 10% [~] 20% of F2 offspring behave the same as parents. Moreover, we found higher levels of defects in chromosome morphology and integrity, and sister-chromatid segregation in F2 population than those in both inbreds and hybrids, which suggest a transgressive effect influencing heat tolerance of meiosis. Our findings reveal that heat resilience in Arabidopsis is controlled by the interplay of multiple genomic loci, holding a great potential to be exploited in crop breeding.

plant biology↗

bioRxiv2024

The asymmetric expression of HSPA2 in blastomeres governs the first embryonic cell-fate decision

The first cell-fate decision is the process by which cells of an embryo take on distinct lineage identities for the first time, thus representing the beginning of developmental patterning. Here, we demonstrate that the molecular chaperone heat shock protein A2 (HSPA2), a member of the 70 kDa heat shock protein (HSP70) family, is asymmetrically expressed in the late 2-cell stage of mouse embryos. The knockdown of Hspa2 in one of the 2-cell blastomeres prevented its progeny predominantly toward the inner cell mass (ICM) fate. In contrast, the overexpression of Hspa2 in one of the two-cell blastomeres did not induce blastomeres to differentiate towards the ICM fate. Furthermore, we demonstrated that HSPA2 interacts with CARM1 and its levels correlate with ICM-associated genes and that it interacts with the CARM1. Collectively, our results identify HSPA2 as a critical early regulator of the first cell-fate decision in mammalian 2-cell embryos.

developmental biology↗

bioRxiv2024

CREEPING STEM1 Regulates Directional Auxin Transport for Lodging Resistance in Soybean

Soybean, a staple crop on a global scale, frequently encounters challenges due to lodging under high planting densities, which results in significant yield losses. Despite extensive research, the fundamental genetic mechanisms governing lodging resistance in soybeans remain elusive. In this study, we identify and characterize the Creeping Stem 1 (CS1) gene, which plays a crucial role in conferring lodging resistance in soybeans. The CS1 gene encodes a HEAT-repeat protein that modulates hypocotyl gravitropism by regulating amyloplast sedimentation. Functional analysis reveals that the loss of CS1 activity disrupts polar auxin transport, vascular bundle development, and the biosynthesis of cellulose and lignin, ultimately leading to premature lodging and aberrant root development. Conversely, increasing CS1 expression significantly enhances lodging resistance and improves yield under conditions of high planting density. Our findings shed light on the genetic mechanisms that underlie lodging resistance in soybeans and highlight the potential of CS1 as a valuable target for genetic engineering to improve crop lodging resistance and yield.

plant biology↗

bioRxiv2024

Network pharmacological evaluation of active ingredients and potential targets of San Huang Decoction in the treatment of breast cancer

ObjectiveUsing the method of network pharmacology, the possible targets and related signaling pathways of San Huang Decoction against breast cancer were explored. MethodsThe active ingredients of San Huang Decoction were screened in Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP). All targets of San Huang Decoction and breast cancer were obtained through five databases, and R software was used to obtain their common targets. Then, Cytoscape software was used to construct the active ingredient-target network map. The protein-protein interaction network (PPI) was obtained with the help of the STRING database, and key targets were screened out through network topology analysis. Common targets were subjected to gene ontology (GO) and Kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis using R software. ResultsA total of 39 active ingredients were screened from San Huang Decoction, of which 20 belonged to Huangqi, 16 belonged to Dahuang, and 3 belonged to Jianghuang. After taking the intersection of all the targets of San Huang Decoction and breast cancer, a total of 185 potential targets of San Huang Decoction against breast cancer were obtained. PPI network analysis screened out 16 targets that may be the key targets of San Huang Decoction against breast cancer. The enrichment analysis of GO and KEGG suggested that the common targets were involved in various biological processes and functions through PI3K-Akt signaling pathway, MAPK signaling pathway, TNF signaling pathway, IL-17 signaling pathway, and HIF-1 signaling pathway, including oxidation stress, transcription-related functions, and nuclear receptor functions. ConclusionThis study predicted the potential targets and signaling pathways of San Huang Decoction against breast cancer, and provided scientific guidance for the further study on the anti-breast cancer mechanism of San Huang Decoction.

bioinformatics↗

bioRxiv2024

Mapping Individual Differences in the Topological Landscape of Naturalistic Brain Dynamics

Psychiatry seeks to unravel brain dysfunction and individual differences in real-world contexts. Naturalistic stimuli, like movie watching, are increasingly recognized for eliciting complex, context-dependent neural activity with high ecological validity. Yet, current methods often rely on standard paradigms that average data across time, limiting the full potential of such stimuli. Here, we present STIM, a Topological Data Analysis-based framework designed to dynamically track how individuals integrate complex contexts in real time. Applied to large-sample fMRI data from movie watching, STIM constructs a robust low-dimensional dynamical landscape that reflects group consensus while probing individual variations at both global (spanning narratives) and local (within specific narratives) levels. At the global level, individual differences emerge along a center-periphery gradient in the dynamical landscape, which significantly predicts fluid intelligence, underscoring the importance of neural adaptability and diversity. At finer scales, local geometric features correlate with context-specific psychological traits beyond cognition. STIM also captures developmental changes in the dynamical landscape and reveals abnormalities in conditions such as autism. These findings demonstrate that STIM leverages the rich information from movie stimuli and fMRI recordings as neural probes to assess individual differences in cognition and mental health.

neuroscience↗

bioRxiv2024

Lipid droplets restrict phagosome formation during Candida challenge

Lipid droplets (LDs) are intracellular organelles which can be induced and interact with phagosomes during the process of pathogen phagocytosis in macrophages. However, the function of LDs in the phagocytosis remains elusive. Here, we unveil the role of LDs in modulating phagosome formation using a fungal infection model. Specifically, LDs accumulation restricted the quantity of phagosome formation, and protected macrophages from death. Mechanistically, LDs formation competitively consumed intracellular endoplasmic reticulum membrane and altered RAC1 translocation and its GTPase activity, which resulted in limiting phagosomes formation in macrophages during fungi engulfed. Mice with Hilpda-deficient macrophages were more susceptible to the lethal sequelae of systemic infection with C. albicans. Notably, administration of ATGL inhibitor Atglistatin improved host outcomes in disseminated fungal infections. Taken together, our study elucidates the mechanism of LDs in controlling phagosomes formation to avoid immune cell death, and offers a potential drug target for treatment of C. albicans infections. HighlightsO_LIC. albicans infection induces triglycerides production and LD accumulation in macrophages. C_LIO_LILD restricts the quantity of phagosome formation and protects macrophages from death. C_LIO_LILD consumes intracellular ER membrane components and alters RAC1 translocation and its GTPase activity. C_LIO_LIAdministration of ATGL inhibitor Atglistatin improves host outcomes in disseminated fungal infections. C_LI

microbiology↗

bioRxiv2024

Genetic burden of dysregulated cytoskeletal organisation in the pathogenesis of pulmonary fibrosis

BackgroundRare genetic variants contribute to pulmonary fibrosis (PF) risk and outcome, with known variants highlighting the importance of impaired telomere maintenance and surfactant biology. However, much of the disrupted genetic architecture of PF remains unexplained. This study aimed to identify genes with rare pathogenic coding variants that represented a burden at the exon level associated with the pathogenesis of PF. MethodsPF case cohorts included the PROFILE study of incident idiopathic pulmonary fibrosis (IPF) and PF-classified participants from the Genomics England 100K (GE100KGP) study. Whole genome sequencing data were used to test the burden of rare protein altering variants (PAVs; CADD score >20) defined by minor allele frequency <0.1% in the summary level gnomAD reference database (v3.1.2). A summary exon-level pathogenicity score was derived by standardising PAV predictions from functional annotation tools (AlphaMissense, REVEL, ClinPred, CADD, PolyPhen-2, SIFT) averaged across exons. SKAT-O based kernel regression associated exon-level pathogenicity features with disease progression and survival. Single cell lung transcriptomic datasets were harmonised to evaluate expression patterns and gene ontology characterised enriched molecular functions. ResultsAcross 507 PROFILE cases and 451 GE100KGP cases, each compared against 76,156 reference participants, 77 genes showed overlapping significant rare PAV burden, comprising 206 concordant PAVs. Of these, 15 genes had exons with pathogenicity scores that were associated with worse clinical outcomes in IPF, including FAT4 exon-10 and DNAH7 exon-43. Dysregulated gene expression of COL6A3 and FAT4 was observed in IPF lung fibroblasts, while DNAH7, DNAH12 and PCM1 showed high expression in IPF epithelial cells. The top enriched molecular functions were related to cytoskeletal motor activity. ConclusionsRare PAV burden highlighted that pathogenicity within specific exons was associated with PF risk and worse clinical outcomes, identifying genes with dysregulated expression in disease lung cells. The genetic architecture implicates disrupted cytoskeletal organisation in the pathogenesis of PF.

genetics↗

bioRxiv2024

Defining cellular determinants of resistance to PD-1 pathway blockade in non-small-cell lung cancer

Despite sustained clinical responses to immune-checkpoint blockade (ICB) therapies in non-small-cell lung cancer (NSCLC), the majority of patients derive no clinical benefits, and the cellular and molecular underpinnings of such resistance remain incompletely understood. To identify cell types that may influence immunotherapy responses, we first integrated newly generated and previously published single-cell RNA sequencing data from 110 treatment-naive patients with NSCLC. Among tumor-resident cell types, we identified MMP1+ cancer-associated fibroblasts (CAFs), which were inversely correlated with the level of tumor-reactive T cells--a key determinant of response to ICB. Further single-cell analysis for newly collected 21 tumor samples from NSCLC patients treated with anti-PD-1/PD-L1 agents revealed that MMP1+ fibroblasts were indeed enriched in treatment-refractory patients, and this observation was also validated in an independent dataset of bulk RNA sequencing from 344 NSCLC patients treated with PD-L1 agents. Examination of the spatial architecture showed that MMP1+ fibroblasts were located at the tumor-stroma boundary, forming a single-cell layer that encircled the cancer cell aggregates, and we hence defined MMP1+ fibroblasts as tumor-stroma boundary (tsb)CAFs. Such tsbCAFs likely promote resistance to ICB by functioning as a physical barrier that prevents tumor-reactive T cells from recognizing and killing cancer cells. Our study provides a new framework to identify cellular underpinnings of resistance to ICB and suggests new strategies to overcome ICB resistance. Highlights{lozenge} Identification and characterization of MMP1+ fibroblasts in lung cancer. {lozenge}Single-cell meta-analysis reveals cell populations impeding the accumulation of tumor-reactive T cells. {lozenge}MMP1+ fibroblasts correlate with the low infiltration of tumor-reactive T cells and the resistance to anti-PD-1/PD-L1 treatment. {lozenge}MMP1+ fibroblasts appear to form a space barrier between malignant and T cells.

immunology↗

bioRxiv2024

Structure of Geobacillus stearothermophilus Cas9: insights into the catalytic process and thermostability of CRISPR-Cas9

CRISPR-Cas9 has been developed as a powerful gene editing tool, but the mechanism governing the intricate catalytic process remains incompletely resolved. Here, the cryo-electron microscopy structures of thermostable Cas9 from Geobacillus stearothermophilus (GeoCas9) in complex with sgRNA and target DNA are reported. The structure of GeoCas9 in complex with sgRNA reveals a slit termed L1-crevice comprising HNH, RuvC, and L1 helix as a transient storage site of 5 spacer of sgRNA. When 5 spacer is extracted to pair with the target DNA, L1-crevice collapses to trigger the subsequent HNH domain translocation. In addition, structural and biochemical analyses suggest that the resilience of GeoCas9 at elevated temperature is related to the unique PI domain conformation. These results advance our understanding into the catalytic process of Cas9 and unveil the molecular mechanism that accounts for the superior thermal profile of GeoCas9.

molecular biology↗

bioRxiv2024

Post-flowering photoperiod sensitivity of soybean in pod-setting responses

The development of soybean (Glycine max) is regulated by photoperiod, with genes related to photoperiod sensitivity primarily focused on flowering time. However, their roles in post-flowering reproductive development and the mechanisms by which photoperiod affects them are not yet determined. In this study, we found that pod formation is sensitive to photoperiod. Long-day (LD) conditions tend to extend the time from flowering to pod formation (R1 to R3 stage), and the first wave of flowers tends to fall off. Additionally, photoperiod affects pistil morphology; under short-day (SD) conditions, the stigma has a curved hook-like structure that facilitates better interaction with the filaments when pollen is released, ultimately influencing the timing of pod formation. Photoperiod-insensitive mutants, lacking E1 family and Evening Complex genes, showed no difference in pod formation time under LD or SD conditions. Hormone content analysis and transcriptome data analysis indicated that various hormones, ROS signals, and the application of sucrose solution in vitro might influence floral organ abscission. HighlightPhotoperiod sensitivity after flowering affects the pod-setting time in soybean.

plant biology↗

bioRxiv2024

Ex Vivo Cortical Circuits Learn to Predict and Spontaneously Replay Temporal Patterns

It has been proposed that prediction and timing are computational primitives of neocortical microcircuits, specifically, that neural mechanisms are in place to allow neocortical circuits to autonomously learn the temporal structure of external stimuli and generate internal predictions. To test this hypothesis, we trained cortical organotypic slices on two specific temporal patterns using dual-optical stimulation. After 24-hours of training, whole-cell recordings revealed network dynamics consistent with training-specific timed prediction. Unexpectedly, there was replay of the learned temporal structure during spontaneous activity. Furthermore, some neurons exhibited timed prediction errors. Mechanistically our results indicate that learning relied in part on asymmetric connectivity between distinct neuronal ensembles with temporally-ordered activation. These findings further suggest that local cortical microcircuits are intrinsically capable of learning temporal information and generating predictions, and that the learning rules underlying temporal learning and spontaneous replay can be intrinsic to local cortical microcircuits and not necessarily dependent on top-down interactions.

neuroscience↗

bioRxiv2024

Single-cell landscapes of long non-coding RNAs in early vascular endothelial development and hemogenic specification

Understanding the molecular regulation of arterial and hemogenic specification during early embryonic vascular development is crucial for guiding vascular and hematopoietic regeneration. Accumulating evidence emphasizes the role of long non-coding RNAs (lncRNAs) in cell fate decision. However, the dynamic expression and the potential roles of lncRNAs in early vascular development are still unknown. Here, we first constructed a single-cell landscape of lncRNA expression based on the deeply sequenced tag-based single-cell transcriptome data of early embryonic vascular endothelial cells (VECs). We revealed the contribution of lncRNAs to VEC heterogeneity and identified 295 lncRNAs with specific expression in eight VEC populations. Furthermore, we identified a series of lncRNAs potentially involved in regulating the two waves of arterial specification and hemogenic specification. We uncovered a transient downregulation of H19 in the hemogenic endothelial population during endothelial-to-hematopoietic transition. Additionally, we constructed a transcription factor regulatory network composed of 287 regulons for early VEC development. We further revealed differential activation patterns of regulons and modules in the eight VEC populations, and predicted potential lncRNA-regulon regulatory network. Moreover, unsupervised analysis of the lncRNA expression profile revealed novel VEC subpopulations strongly associated with the maturation of VECs, suggesting the prominent roles of lncRNAs in endothelial maturation. In summary, our study fills the gap in understanding of lncRNA regulatory networks in early vascular development and provides insights into the fields of vascular and hematopoietic regeneration research.

developmental biology↗

bioRxiv2024

Cell-type specific inference from bulk RNA-sequencing data by integrating single cell reference profiles via EPIC-unmix

Cell type specific (CTS) analysis is essential to reveal biological insights obscured in bulk tissue data. However, single-cell (sc) or single-nuclei (sn) resolution data are still cost-prohibitive for large-scale samples. Thus, computational methods to perform deconvolution from bulk tissue data are highly valuable. We here present EPIC-unmix, a novel two-step empirical Bayesian method integrating reference sc/sn RNA-seq data and bulk RNA-seq data from target samples to enhance the accuracy of CTS inference. We demonstrate through comprehensive simulations across three tissues that EPIC-unmix achieved 4.6% - 109.8% higher accuracy compared to alternative methods. By applying EPIC-unmix to human bulk brain RNA-seq data from the ROSMAP and MSBB cohorts, we identified multiple genes differentially expressed between Alzheimers disease (AD) cases versus controls in a CTS manner, including 57.4% novel genes not identified using similar sample size sc/snRNA-seq data, indicating the power of our in-silico approach. Among the 6-69% overlapping, 83%-100% are in consistent direction with those from sc/snRNA-seq data, supporting the reliability of our findings. EPIC-unmix inferred CTS expression profiles similarly empowers CTS eQTL analysis. Among the novel eQTLs, we highlight a microglia eQTL for AD risk gene AP3B2, obscured in bulk and missed by sc/snRNA-seq based eQTL analysis. The variant resides in a microglia-specific cCRE, forming chromatin loop with AP3B2 promoter region in microglia. Taken together, we believe EPIC-unmix will be a valuable tool to enable more powerful CTS analysis.

genomics↗

bioRxiv2024

Gene2role: a role-based gene embedding method for comparative analysis of signed gene regulatory networks

MotivationUnderstanding the dynamics of gene regulatory networks (GRNs) across various cellular states is crucial for deciphering the underlying mechanisms governing cell behavior and functionality. However, current comparative analytical methods, which often focus on simple topological information such as the degree of genes, are limited in their ability to fully capture the similarities and differences among the complex GRNs. ResultsWe present Gene2role, a gene embedding approach that leverages multi-hop topological information from genes within signed GRNs. Initially, we demonstrated the effectiveness of Gene2role in capturing the intricate topological nuances of genes using GRNs inferred from four distinct data sources. Then, applying Gene2role to integrated GRNs allowed us to identify genes with significant topological changes across cell types or states, offering a fresh perspective beyond traditional differential gene expression analysis. Additionally, we quantified the stability of gene modules between two cellular states by measuring the changes in the gene embeddings within these modules. In conclusion, our method augments the existing toolkit for probing the dynamic regulatory landscape, thereby opening new avenues for understanding gene behavior and interaction patterns across cellular transitions.

bioinformatics↗

bioRxiv2024

Multi-omics reveals AKR1B1-regulated galactose metabolic as a driver of gastrointestinal stromal tumor progression

The underlying mechanism of malignant progression in gastrointestinal stromal tumors (GISTs) is not fully understood. Despite recent advancements, a comprehensive profile of metabolome, transcriptome, and proteome of GISTs is lacking. This study conducted an integrated multi-omics analysis of GISTs across different risk classifications. By integrating metabolomics, transcriptomics, and proteomics, we identify distinct metabolic patterns and associated biological pathways implicated in the malignant progression of GISTs. Moreover, we identified galactose metabolism and the pivotal rate-limiting enzyme AKR1B1 is dysregulated in GISTs progression. AKR1B1 was upregulated and predicted poor prognosis in GISTs. In addition, AKR1B1 knockdown resulted in trehalose accumulation in GIST cells, thereby inhibiting cell proliferation and mitosis. These findings not only enhance our comprehension of the underlying mechanisms governing GIST progression from a metabolic reprogramming standpoint but also furnish prognostic biomarkers and potential therapeutic targets for GISTs.

cancer biology↗

bioRxiv2024

Quantitative and systematic NMR measurements of sequence-dependent A-T Hoogsteen dynamics uncovers unique conformational specificity in the DNA double helix

The propensities to form lowly-populated short-lived conformations of DNA could vary with sequence, providing an important source of sequence-specificity in biochemical reactions. However, comprehensively measuring how these dynamics vary with sequence is challenging. Using 1H CEST and 13C R1{rho} NMR, we measured Watson-Crick to Hoogsteen dynamics for an A-T base pair in thirteen trinucleotide sequence contexts. The Hoogsteen population and exchange rate varied 4-fold and 16-fold, respectively, and were dependent on both the 3- and 5-neighbors but only weakly dependent on monovalent ion concentration (25 versus 100 mM NaCl) and pH (6.8 versus 8.0). Flexible TA and CA dinucleotide steps exhibited the highest Hoogsteen populations, and their kinetics rates strongly depended on the 3-neighbor. In contrast, the stiffer AA and GA steps had the lowest Hoogsteen population, and their kinetics were weakly dependent on the 3-neighbor. The Hoogsteen lifetime was especially short when G-C neighbors flanked the A-T base pair. The Hoogsteen dynamics had a distinct sequence-dependence compared to duplex stability and minor groove width. Thus, our results uncover a unique source of sequence-specificity hidden within the DNA double helix in the form of A-T Hoogsteen dynamics and establish the utility of 1H CEST to quantitively measure sequence-dependent DNA dynamics.

biochemistry↗

bioRxiv2024

TMEM165 replenishes lysosomal calcium stores, protects cells against calcium overload, and mediates calcium-induced lysosomal H+ leakage

The proper function of lysosomes depends on their ability to store and release calcium. While several lysosomal calcium release channels have been described, how mammalian lysosomes replenish their calcium stores has not been determined. Using genetic depletion and overexpression techniques combined with electrophysiology and visualization of subcellular ion concentrations and their fluxes across the lysosomal membrane, we show here that TMEM165 imports calcium to the lysosomal lumen and mediates calcium-induced lysosomal proton leakage. Accordingly, TMEM165 accelerates the recovery of cells from cytosolic calcium overload thereby enhancing cell survival while causing a significant acidification of the perilysosomal area and the entire cytosol. These data indicate that in addition to its essential role in the glycosylation in the Golgi, a small but significant fraction of TMEM165 localizes on the lysosomal limiting membrane, where it protects cells against cytosolic calcium overload, preserves lysosomal function by refilling lysosomal calcium stores and regulates perilysosomal and lysosomal pH.

cell biology↗

bioRxiv2024

FAPM: Functional Annotation of Proteins using Multi-Modal Models Beyond Structural Modeling

Assigning accurate property labels to proteins, like functional terms and catalytic activity, is challenging, especially for proteins without homologs and "tail labels" with few known examples. Unlike previous methods that mainly focused on protein sequence features, we use a pretrained large natural language model to understand the semantic meaning of protein labels. Specifically, we introduce FAPM, a contrastive multi-modal model that links natural language with protein sequence language. This model combines a pretrained protein sequence model with a pretrained large language model to generate labels, such as Gene Ontology (GO) functional terms and catalytic activity predictions, in natural language. Our results show that FAPM excels in understanding protein properties, outperforming models based solely on protein sequences or structures. It achieves state-of-the-art performance on public benchmarks and in-house experimentally annotated phage proteins, which often have few known homologs. Additionally, FAPMs flexibility allows it to incorporate extra text prompts, like taxonomy information, enhancing both its predictive performance and explainability. This novel approach offers a promising alternative to current methods that rely on multiple sequence alignment for protein annotation. The online demo is at: https://huggingface.co/spaces/wenkai/FAPM_demo.

bioinformatics↗