Biology subjects

Ding, F.

Publications and source records attributed to Ding, F..

At least 19 recordsLinked to original sources

bioRxiv2026

Single-probe, high-yield spatial mapping of diverse short RNA motifs

Spatial transcriptomics has revolutionized our understanding of RNA biology. Hybridization-based methods map mRNA transcripts by tiling multiple labeled probes along each molecule to generate a visible fluorescent signal, restricting applicability to long transcripts. Short RNA motifs, consequently, can only accommodate a single probe and remain a spatially invisible dimension of RNA biology. Here we present SMOOTHY-FISH, a single probe strategy that directly visualizes short RNA motifs at the single-molecule resolution. By optimizing probe brightness, binding specificity, and target accessibility, our single ~20-nucleotide probe reached the practical ceiling of hybridization-based detection. We quantitatively mapped previously inaccessible short motifs - splice-isoform junctions, individual microRNAs, and structural conformations - in single cells and mouse brain tissues, each validated against orthogonal methods. This unveiled biological phenomena that were previously unknown: ratios of alternatively spliced isoforms are tightly regulated across cell populations, subcellular miRNA fractions are precisely maintained, and a designed RNA structure was found to have a prominent alternative conformation. These findings demonstrate that short RNA motifs possess key spatial information that has yet to be discovered. With SMOOTHY-FISH's optimized single probe technique, short RNA motifs are converted into a new routinely measurable axis of spatial transcriptomics.

molecular biology↗

bioRxiv2026

The multiscale architecture of trait aggression: from meta-analytic convergence to multimodal characterization

Background. Trait aggression shows substantial interindividual variability, yet its neurobiological architecture across structural, functional, molecular, and macroscale organizational levels remains poorly understood. Methods. We developed a convergent multiscale framework to characterize the neurobiological architecture associated with trait aggression. We first integrated a large-sample voxel-based morphometry analysis (n = 922) with published neuroimaging studies using two complementary coordinate-based meta-analytic approaches, activation likelihood estimation and seed-based d mapping (total N = 3,671). Convergent meta-analytic findings were subsequently mapped onto functional connectivity networks, and their molecular signatures were characterized using imaging transcriptomics and neurotransmitter mapping. Finally, we examined whether the spatial organization of aggression-related networks was embedded within intrinsic multimodal brain hierarchies spanning morphometric, functional, transcriptomic, and neurochemical dimensions. Results. Across complementary analyses, trait aggression converged on a reproducible structural-functional network centered on the dorsomedial prefrontal cortex and dorsal anterior cingulate cortex. This architecture exhibited distinct molecular signatures, including gene-expression patterns enriched for synaptic functions and robust spatial associations with serotonergic, dopaminergic, and noradrenergic transporter systems. Moreover, the spatial organization of the aggression-related functional network was systematically predicted by the brain's intrinsic multimodal gradients, indicating that it is embedded within broader principles of cortical hierarchical organization. Conclusions. these findings move beyond the localization of aggression to discrete brain regions and identify a reproducible multiscale architecture linking aggression-related structural variation to functional network organization, molecular signatures, and the brain's intrinsic multimodal hierarchy. This integrative framework provides a systems-level account of the neurobiological basis of individual differences in trait aggression.

neuroscience↗

bioRxiv2026

Functional and evolutionary insights into the emerging tet(X4)-carrying non-O1/O139 Vibrio cholerae from retail pork

The tigecycline resistance gene tet(X4) is prevalent in Enterobacteriaceae, particularly in Escherichia coli. To our knowledge, no study has reported the dissemination dynamics of tet(X4) in Vibrio spp. Herein, we isolated and characterized a first tet(X4)-positive non-O1/O139 Vibrio cholerae isolate from retail pork. Genomic sequencing identified a novel tet(X4) variant in the V. cholerae chromosome, harboring a G568A nucleotide substitution that resulted in an Ala190Thr (A190T) amino acid substitution in Tet(X4). While this Tet(X4)-A190T variant conferred lower phenotypic resistance to tetracyclines (including tigecycline) than the wild-type Tet(X4), its overall catalytic efficiency against these antibiotics was paradoxically enhanced despite a reduced substrate affinity. Genomic comparisons revealed that two copies of ISCR2 flanked the variant gene, and the structure was ISCR2-hp-hp-abh-tet(X4)G568A -ISCR2, which is highly homologous to the reported E. coli plasmids carrying tet(X4). In addition, it confirmed the presence of an ISCR2-mediated circular intermediate, proving this modules capacity for horizontal transfer of the tet(X4)G568A variant. Furthermore, the ISCR2-tet(X4) genetic structure carrying the G568A substitution was integrated within a chimeric SXT/R391-like integrative and conjugative element (ICE), which is also serving as a vehicle for genetic dissemination. As per our knowledge, this is the first report on the emergence of SXT/R391-like ICE carrying tet(X4) in Vibrio strains. Our finding demonstrates that the clinically relevant tigecycline resistance gene tet(X4), previously confined mainly to Enterobacterales from humans and livestock, is now actively spreading into environmental Vibrio populations. This cross-species transfer highlights a previously underappreciated ecological and public health concern in aquatic ecosystems.

microbiology↗

bioRxiv2026

NRF2 pathway activation and SPP1⁺TREM2⁺ macrophages drive chemoradiotherapy resistance in esophageal squamous cell carcinoma

Esophageal squamous cell carcinoma (ESCC) is among the most aggressive cancers, with low rates of durable response to chemoradiotherapy and limited therapeutic options for relapsed disease. To uncover mechanisms of treatment resistance and relapse, we performed comprehensive multi-omics profiling of >100 pre-treatment and post-relapse ESCC tumors from a prospective clinical trial (NCT04694391), integrating whole-exome/genome sequencing, bulk RNA-sequencing, single-cell RNA sequencing, and spatial transcriptomics. We identify somatic alterations in NFE2L2/KEAP1 in nearly 40% of relapsed patients, which are associated with upregulation of NRF2 signaling targets in resistant tumors and cell line models. At single-cell and spatial resolution, relapsed tumors are enriched for NRF2-activated epithelial cells that physically co-localize with immunosuppressive SPP1TREM2 macrophages. This co-localization suggests a synergistic interaction between NRF2-driven tumor programs and macrophage-mediated immune suppression that promotes relapse after chemoradiotherapy. Our findings nominate NFE2L2/KEAP1 mutations as predictive biomarkers for patient stratification and highlight therapeutic targeting of NRF2 signaling and SPP1TREM2 macrophages as rational strategies to overcome resistance in ESCC.

cancer biology↗

bioRxiv2025

Sodium Selenite Alleviates Hyperglycemia-aggravated Cerebral Ischemic Injury by Mediating Apoptosis through the Fas/FasL Signaling Pathway

ObjectiveTo investigate the intervention mechanism of sodium selenite in alleviating hyperglycemia-induced cerebral ischemia/reperfusion (I/R) injury through the Fas/FasL/NF-{kappa}B/PUMA pathway. MethodsThe SD rat model of hyperglycemia-induced cerebral I/R injury and the HT22 cell model of high-glucose oxygen deprivation/reoxygenation (OD/R) were constructed to systematically evaluate the neuroprotective effect of sodium selenite and its regulatory effect on the Fas/FasL/NF-{kappa}B/PUMA pathway. ResultsThe infarct volume, neurological dysfunction, metabolic level, and apoptosis level in the HG group were significantly higher than those in the NG group. Sodium selenite improved the infarct volume, neurological dysfunction, metabolic level, and apoptosis level in the HG group, and regulated the Fas/FasL/PUMA apoptotic pathway to alleviate injury. In vitro experiments showed that sodium selenite significantly improved morphological damage caused by high-glucose OD/R, increased cell viability, and reduced the expression of apoptotic factors. The application of Fas and FasL inhibitors and knockdown of PUMA expression indicated that sodium selenite alleviated cell apoptosis by inhibiting the activation of the Fas/FasL/PUMA pathway. ConclusionSodium selenite alleviates hyperglycemia-exacerbated cerebral I/R injury by regulating the Fas/FasL/NF-{kappa}B/PUMA signaling pathway.

pathology↗

bioRxiv2025

A Global Thermodynamic-Kinetic Model Capturing the Hallmarks of Liquid-Liquid Phase Separation and Amyloid Aggregation

Aberrant aggregation of proteins into amyloid fibrils is associated with numerous neurodegenerative, systemic and metabolic diseases. Amyloidogenic proteins undergo spontaneous liquid-liquid phase separation (LLPS), rapidly forming protein-rich condensates prior to fibrillization. However, the exact effects of LLPS on amyloid aggregation remain unclear as contrasting fibrillization-promotion, inhibition and even biphasic effects have been reported in the literature. In this study, we integrate LLPS-induced heterogeneity of protein concentrations into a thermodynamic-kinetic model of amyloid aggregation. We adopt the phase transition theory and introduce protein condensates as an additional protein state alongside non-interacting monomers, oligomers and fibrils. Oligomerization and fibrillization can occur both in the protein-rich condensates and the protein-poor solution. This model allows us to derive the time evolution of different states - monomers, condensates, oligomers, and fibrils - spanning a wide range of concentrations, and determine how model parameters related to LLPS, fibrillization, and oligomerization influence fibrillization kinetics. Using this global model, we resolve the seemingly contradictory effects of LLPS on fibrillization. We expect the developed thermodynamic-kinetic model of LLPS, and amyloid aggregation will help advance our understanding, modulation, and mitigation of pathological aggregation processes in amyloid diseases. TOC O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=62 SRC="FIGDIR/small/655849v1_ufig1.gif" ALT="Figure 1"> View larger version (13K): org.highwire.dtl.DTLVardef@12eeb58org.highwire.dtl.DTLVardef@15cee66org.highwire.dtl.DTLVardef@278c10org.highwire.dtl.DTLVardef@5c8d4a_HPS_FORMAT_FIGEXP M_FIG C_FIG

biophysics↗

bioRxiv2025

Synthesizing human-specific spike-in standards for RNA-seq experiments and assessing their technical performance

Using spike-in standards for RNA-seq experiments is critical to evaluate technical bias introduced during sample preparation, sequencing and analysis. Although some external RNA spike-in standards have been developed, species-specific spike-in standard was not reported yet. Here we developed the human-specific spike-in standards with 65 controls. We first extracted human endogenous RNAs with various lengths and GC contents, introduced random mutations approximately every 75 bp in each RNA, and then synthesized these RNAs as spike-in RNAs. After that, four mixtures of these spike-in RNAs covering a 220 dynamic range were obtained. To ensure the accuracy of RNA concentration, two rounds of ddPCR were conducted for each spike-in RNA and the intraclass correlation coefficient between two ddPCRs ranged from 0.9954 to 0.9971 after removing the two spike-in RNAs with the largest concentration difference. Furthermore, we showed that the sequencing error profiles were distinct between platforms and the library preparation procedures were related with the discrepancies in spike-in RNA read percent, transcript abundance, sequence coverage distribution, and differential gene expression. In addition, two regression models of sequence coverage were built based on RNA second structure and GC content, and 86.62%-91.78% of the variation can be explained. Our study demonstrates the technical performance of the human-specific spike-in standards for RNA-seq experiments and illustrates the biases across different libraries, platforms, and laboratories.

bioinformatics↗

bioRxiv2025

Protein Barcoding and Next-Generation Protein Sequencing for Multiplexed Protein Selection, Analysis, and Tracking

Protein barcoding has emerged as a powerful tool for the multiplexed identification and characterization of proteins, providing a mechanism for precise tracking of protein affinity, location, and expression. In this study, we describe the development of a protein barcoding workflow for use with single-molecule Next-Generation Protein Sequencing (NGPS) on the benchtop Platinum(R) and Platinum(R) Pro instruments. We present data on the validation of eight peptide barcodes, each designed to minimize detection bias and maximize sensitivity across various experimental conditions. We have also optimized the design of expression constructs to decrease both the hands-on time and input requirements of the workflow. In this workflow, affinity-tagged proteins are expressed with unique peptide barcodes. Following experimental selection or treatments, the proteins are purified, and the peptide barcodes are cleaved and sequenced on the Platinum instrument. We demonstrate that we can detect barcodes at 400 fmol of sample input concentration within the eight-plex mixture, and at 50 fmol of sample input for individual barcodes. We also show the capacity of this barcoding approach to achieve a ten-fold dynamic range, underscoring its sensitivity in recovering variants with low abundance. Through the combination of protein barcoding and NGPS, we lay the groundwork for future studies aimed at characterizing protein interactions and improving targeted drug delivery strategies.

biochemistry↗

bioRxiv2024

Exploration of the Roles of HLAs When Predicting Infection Status by T Cell Receptors

T cells are critical components of human immune system. When a cell is infected by a virus, it presents viral peptides on its surface using human leukocyte antigen (HLA) proteins. These peptide-HLA complexes are recognized by T cells through interactions with T cell receptors (TCRs). A human blood sample can contain millions of unique TCRs, which is a sample from the TCR repertoire of the individual. TCR repertoire-wide association studies (TReWAS) aim to evaluate the association between individual TCRs and disease or exposure status. Previous studies have demonstrated that TCRs associated with viral infections can be identified through TReWAS, and such TCRs can be used to predict current or past infection with high accuracy. Notably, many TCRs are strongly associated with specific HLA alleles, suggesting that incorporating HLA information could enhance the accuracy of TReWAS analyses and TCR-based predictions. In our study, we evaluated TCR-based predictions by conditioning on individual HLA alleles or their k-nearest neighbors. We observed improved prediction accuracy for certain HLA alleles. Furthermore, these HLA-specific predictions provide insights into the role of specific HLAs in the infection or disease of interest, offering potential applications in personalized medicine.

bioinformatics↗

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

A nuclear RNA degradation code for eukaryotic transcriptome surveillance

The RNA exosome plays critical roles in eukaryotic RNA degradation, but it remains unclear how the exosome specifically recognizes its targets. The PAXT connection is an adaptor that recruits the exosome to polyadenylated RNAs in the nucleus, especially transcripts polyadenylated at intronic poly(A) sites. Here we show that PAXT-mediated RNA degradation is induced by the combination of a 5' splice site and a poly(A) junction, but not by either sequence alone. These sequences are bound by U1 snRNP and cleavage/polyadenylation factors, which in turn cooperatively recruit PAXT. As the 5' splice site-poly(A) junction combination is typically not found on correctly processed full-length RNAs, we propose that it functions as a "nuclear RNA degradation code" (NRDC). Importantly, disease-associated single nucleotide polymorphisms that create novel 5' splice sites in 3' untranslated regions can induce aberrant mRNA degradation via the NRDC mechanism. Together our study identified the first NRDC, revealed its recognition mechanism, and characterized its role in human diseases.

molecular biology↗

bioRxiv2024

APOE regulates the transport of GM1

Apolipoprotein E (APOE) is responsible for lipid transport, including cholesterol transport and clearance. While the {varepsilon}4 allele of APOE (APOE4) is associated with a significant genetic risk factor for late-onset Alzheimers disease (AD), no mechanistic understanding of its contribution to AD etiology has been established yet. In addition to cholesterol, monosialotetrahexosylganglioside (GM1) is a crucial lipid component in cell membranes and has been implicated in promoting the aggregation of amyloid beta protein (A{beta}), a key protein associated with AD. Here, we ask whether there are direct interactions between APOE and GM1 that further impact AD pathology. We find that both APOE3 and APOE4 exhibit superior binding affinity to GM1 compared to cholesterol and have an enhanced cellular uptake to GM1 lipid structures than cholesterol lipid structures. APOE regulates the transport process of GM1 depending on the cell type, which is influenced by the expression of APOE receptors in different cell lines and alters GM1 contents in cell membranes. We also find that the presence of GM1 alters the secondary structure of APOE3 and APOE4 and enhances the binding affinity between APOE and its receptor low-density lipoprotein receptor (LDLR), consequently promoting the cellular uptake of lipid structures in the presence of APOE. To understand the enhanced cellular uptake observed in lipid structures containing 20% GM1, we determined the distribution of GM1 on the membrane and found that GM1 clustering in lipid rafts, thereby supporting the physiological interaction between APOE and GM1. Overall, we find that APOE plays a regulatory role in GM1 transport, and the presence of GM1 on the lipid structures influences this transport process. Our studies introduce a plausible direct link between APOE and AD etiology, wherein APOE regulates GM1, which, in turn, promotes A{beta} oligomerization and aggregation.

neuroscience↗

bioRxiv2024

Protein language models are biased by unequal sequence sampling across the tree of life

Protein language models (pLMs) trained on large protein sequence databases have been used to understand disease and design novel proteins. In design tasks, the likelihood of a protein sequence under a pLM is often used as a proxy for protein fitness, so it is critical to understand what signals likelihoods capture. In this work we find that pLM likelihoods unintentionally encode a species bias: likelihoods of protein sequences from certain species are systematically higher, independent of the protein in question. We quantify this bias and show that it arises in large part because of unequal species representation in popular protein sequence databases. We further show that the bias can be detrimental for some protein design applications, such as enhancing thermostability. These results highlight the importance of understanding and curating pLM training data to mitigate biases and improve protein design capabilities in under-explored parts of sequence space.

bioinformatics↗

bioRxiv2022

High-resolution de novo structure prediction from primary sequence

Recent breakthroughs have used deep learning to exploit evolutionary information in multiple sequence alignments (MSAs) to accurately predict protein structures. However, MSAs of homologous proteins are not always available, such as with orphan proteins or fast-evolving proteins like antibodies, and a protein typically folds in a natural setting from its primary amino acid sequence into its three-dimensional structure, suggesting that evolutionary information and MSAs should not be necessary to predict a proteins folded form. Here, we introduce OmegaFold, the first computational method to successfully predict high-resolution protein structure from a single primary sequence alone. Using a new combination of a protein language model that allows us to make predictions from single sequences and a geometry-inspired transformer model trained on protein structures, OmegaFold outperforms RoseTTAFold and achieves similar prediction accuracy to AlphaFold2 on recently released structures. OmegaFold enables accurate predictions on orphan proteins that do not belong to any functionally characterized protein family and antibodies that tend to have noisy MSAs due to fast evolution. Our study fills a much-encountered gap in structure prediction and brings us a step closer to understanding protein folding in nature.

bioinformatics↗

bioRxiv2022

The immunogenomic landscape of cattle

Here, we report the de novo assembly of a cattle genome using ultra-long-read nanopore sequencing in conjunction with other advanced technologies. The assembled genome contains only 145 contigs (N50 [~] 74.0 Mb). Compare to the current reference cattle genome ARS-UCD1.2, 154 gaps are closed, and 467 scaffolds are further placed in our assembly. Importantly, except two remained gaps in the T-cell receptor /{delta} (TRA/TRD) region, the gene loci of other TRs and immunoglobulins (IGs) are seamlessly assembled and exquisitely annotated. With the characterization of 258 IG genes and 626 TR genes that distributed in seven genomic loci, we illustrate the highest immune gene diversity in mammals to our knowledge. Moreover, the gene structures of major histocompatibility complex (MHC) are integrally depicted with properly phased haplotypes. Thus, our work not only reports a cattle genome with the most continuity and completeness, but also provide a comprehensive view of the complex immune- genome.

genomics↗

bioRxiv2022

Displacement and dissociation of oligonucleotidesduring DNA hairpin closure under strain

The simple hybridization kinetic of an oligonucleotide to its template is a fundamental step in many biological processes such as replication arrest, CRISPR recognition, DNA sequencing, DNA origami, etc. Although approaches exist that address special cases of this problem, there are no simple general prediction schemes. In this work, we have measured experimentally, with no fluorescent labelling, the displacement of an oligonucleotide from its substrate in two situations: one corresponding to oligonucleotide binding/unbinding on ssDNA and one where the oligonucleotide is displaced by the refolding of a dsDNA fork. In this second situation, the fork is expelling the oligonucleotide thus reducing significantly its residence time. To account for our data in these two situations, we have constructed a mathematical model, based on the known nearest neighbor dinucleotide free energies, and provided a good estimate of the residence times of different oligonucleotides (DNA, RNA, LNA) with various length in different experimental conditions (force, temperature, buffer conditions, presence of mismatches, etc.). This study provides a foundation for the dynamics of oligonucleotide displacement, a process of importance in a number of biological and bioengineering contexts.

biophysics↗

bioRxiv2022

Cost-efficient boundary-free surface patterning achieves high effective-throughput of time-lapse microscopy experiments

Time-lapse microscopy plays critical roles in the studies of cellular dynamics. However, to set up a time-lapse movie experiments is not only laborious but also with low output, mainly due to the cell-losing problem (i.e., cell moving out of limited field of view), especially in a long time recording. To overcome these issues, we have designed a cost-efficient way that enables cell patterning on the imaging surfaces without any physical boundaries. Using mouse embryonic stem cells as an example system, we have demonstrated that our boundary-free patterned surface solves the cell-losing problem without disturbing their cellular phenotype. Statistically, the presented system increases the effective-throughput of time-lapse microscopy experiments by order of magnitude.

bioengineering↗

bioRxiv2022

Proximal Exploration for Model-guided Protein Sequence Design

Designing protein sequences with a particular biological function is a long-lasting challenge for protein engineering. Recent advances in machine-learning-guided approaches focus on building a surrogate sequence-function model to reduce the burden of expensive in-lab experiments. In this paper, we study the exploration mechanism of model-guided sequence design. We leverage a natural property of protein fitness landscape that a concise set of mutations upon the wild-type sequence are usually sufficient to enhance the desired function. By utilizing this property, we propose Proximal Exploration (PEX) algorithm that prioritizes the evolutionary search for high-fitness mutants with low mutation counts. In addition, we develop a specialized model architecture, called Mutation Factorization Network (MuFacNet), to predict low-order mutational effects, which further improves the sample efficiency of model-guided evolution. In experiments, we extensively evaluate our method on a suite of in-silico protein sequence design tasks and demonstrate substantial improvement over baseline algorithms.

bioinformatics↗