Biology subjects

Yin, R.

Publications and source records attributed to Yin, R..

At least 19 recordsLinked to original sources

bioRxiv2026

A lactate-HIF1-VDR positive feedback loop drives protective SPP1+ macrophage differentiation to inhibit schistosomiasis induced liver fibrosis

Hepatic fibrosis remains a major cause of morbidity and mortality in patients with schistosomiasis without effective therapy. Here we show that vitamin D receptor (VDR) signaling in macrophages is essential for limiting fibrosis, acting through a previously unrecognized metabolic epigenetic circuit. Hepatic macrophages exhibit the highest VDR expression among liver resident cells, and pharmacological VDR activation with paricalcitol selectively expands a protective SPP1+ macrophage subset derived from circulating monocytes. Myeloid specific VDR knockout exacerbates fibrosis and abrogates paricalcitol's hepatoprotective effects, whereas SPP1+ macrophage depletion worsens disease. Mechanistically, VDR activation synergizes with hypoxia and lactate to drive SPP1 expression via glycolytic reprogramming. Moreover, lactate and HIF1alpha cooperatively induce VDR transcription by binding to the Vdr promoter and enhancing histone lactylation, forming a positive feedback loop that amplifies the antifibrotic response. Our findings establish the lactate-HIF1alpha-VDR-SPP1 axis as an endogenous defense mechanism and identify macrophage VDR as a promising therapeutic target for fibrotic liver diseases.

immunology↗

bioRxiv2026

Gene-specific exponent-corrected normalization for library size in bulk RNA-seq

Correcting for library size is an essential step in bulk RNA-seq analyses, as differences in sequencing depth across samples can obscure biological signal with technical noise. While numerous normalization methods and model-based strategies have been proposed, we demonstrate here that library size-normalized counts and differential expression results obtained from such widely adopted approaches often remain strongly correlated with library size in large-scale RNA-seq experiments. Through a systematic analysis of over 100 publicly available GEO and TCGA RNA-seq datasets with raw count data, we show that library size association is observed for a substantial proportion of genes even after state-of-the-art library size correction approaches recommended by leading normalization tools. To address this issue, we propose gecco, a gene-specific exponent-corrected normalization method for RNA-seq counts that incorporates library size directly into the statistical framework via a gene-specific correction term, rather than applying a uniform adjustment factor across all genes. This formulation generalizes existing normalization approaches and yields normalized counts that are free of residual library size effects. Using both simulation studies and real large-scale RNA-seq datasets, we show that our method mitigates library size bias while preserving biological signal across a range of parameter settings. We further demonstrate that our approach leads to higher detection accuracy and more biologically meaningful pathway enrichment results in downstream differential expression and rhythmicity analyses without compromising false discovery rate control. Our method is implemented in R and is fully compatible with the widely used differential expression analysis methods DESeq2 and edgeR.

bioinformatics↗

bioRxiv2026

Benchmarking AlphaFold and related deep learning approaches for modeling antibody and TCR antigen recognition

Determining the structural basis of antigen recognition by antibodies and T cell receptors (TCRs) provides critical insights into effective immune targeting and can inform design of biotherapeutics and vaccines. Accurate computational modeling of antibodies and TCRs in complex with their targets poses a major challenge for predictive methods, including AlphaFold, which is generally accurate for modeling protein complexes but has shown limited success for immune recognition. In this study we assessed the performance of AlphaFold2, AlphaFold3, increased sampling protocols, and related deep learning methods for modeling antibody-protein, antibody-peptide, and TCR-peptide-major histocompatibility complex (pMHC) recognition. We show that increased sampling and AlphaFold3 generally improve performance relative to default sampling and AlphaFold2, however predictive accuracy and improvement levels varied considerably among interface classes, with antibody-peptide complexes representing a challenge despite their small antigen size. Comparing per-case success across methods showed some complementarity, indicating opportunities for increased success through model pooling approaches, for instance increasing antibody-peptide near-native success from 41% to 59%. Analysis of AlphaFold confidence scores and modeling of a noncanonical complex provided further insights into predictive performance. These results highlight considerations for predictive antibody and TCR complex modeling efforts, while revealing key distinctions among protocols, scoring, and immune complex classes.

bioinformatics↗

bioRxiv2026

DamageFormer: a damage-aware multimodal deep learning framework for DNA lesion identification from nanopore sequencing

BackgroundDNA lesions arise from endogenous metabolism and environmental exposure and are the major drivers of mutagenesis, aging, and cancer development. However, mapping DNA damage at nucleotide resolution remains a technically challenging task. Nanopore sequencing enables direct detection of chemical perturbations through alterations in ionic current signals. Despite this potential, existing computational approaches remain limited in their capacity to generalize across diverse lesion types and to effectively integrate nucleotide sequence context with raw signal information for accurate detection and localization. ResultsWe presented DamageFormer, a multimodal deep learning framework for detection and localization of DNA lesions using native nanopore sequencing data. Central to this framework is LesionBERT, a damage-aware genomic foundation model built upon DNABERT-2 and enhanced with lesion-focused reconstruction objectives to improve representation of chemically modified bases. DamageFormer integrated LesionBERT with a neural signal model through an adaptive gating mechanism, enabling dynamic weighting of sequence context and nanopore signal evidence. The model was trained using a joint objective that combines prediction, localization, and contrastive alignment losses to promote cross-modal coherence and spatial precision. On an oxidative DNA damage benchmark comprising paired sequence and signal data, DamageFormer achieved an AUROC of 0.99997 for lesion detection and a mean absolute localization error of 0.00439, consistently outperforming state-of-the-art baselines. Model interpretation analyses revealed context-dependent modality weighting that adapts to variation in signal quality and sequence ambiguity. The proposed framework further generalizes to chemically distinct guanine lesions not observed during the training process, demonstrating its robustness and transferability to unseen damage types. ConclusionsDamage-aware biological language modeling combined with adaptive multimodal fusion enables accurate and interpretable identification of DNA lesions from nanopore sequencing data. This framework provides a scalable approach for characterizing genome-wide damage landscapes and illustrates how chemical DNA information can be systematically incorporated into genomic language models. The source code and pretrained models of this work are available at: https://github.com/UF-HOBIYin-Lab/DamageFormer.

genomics↗

bioRxiv2026

ST-PARM: Pareto-Complete Inference-Time Alignment for Multi-Objective Protein Design

MotivationProtein engineering is inherently multi-objective: improving one property can degrade others, so practical workflows require generating non-dominated (Pareto-optimal) candidates spanning a trade-off surface. Linear objective scalarization and deterministic pairwise preference learning can under-explore non-convex Pareto regions and amplify noise from uncertain evaluators, limiting Pareto coverage and trade-off controllability. ResultsWe introduce Smooth Tchebycheff Preference-Aware Reward Model (ST-PARM), an inference-time alignment framework that steers a frozen protein language model along user-specified trade-offs with a lightweight reward model trained only once. ST-PARM combines (i) a reward-calibrated pairwise preference loss that is uncertainty-aware by down-weighting ambiguous comparisons under noisy evaluators, (ii) a smooth Tchebycheff scalarization that is Pareto-complete in principle and improves empirical trade-off coverage, and (iii) latent-space pair-construction strategies. On GFP fluorescence-stability (full-length design) and IL-6 nanobody stability-solubility (CDR3+suffix design), ST-PARM delivers broader Pareto coverage and stronger preference tracking than baselines PARM and MosPro. For GFP, a conservative structural screen for local confidence and global fold preservation retains a broad frontier and strong controllability, yielding an actionable cohort for downstream assays. We also provide cross-evaluator robustness checks, a three-objective extension, and a natural-language alignment generality check in the Supplement, establishing a practical foundation for controllable sequence generation under competing multi-objectives and noisy measurements. Availability and Implementationhttps://github.com/Shen-Lab/ST-PARM. Supplementary InformationSupplementary data are provided with the submission.

bioinformatics↗

bioRxiv2026

Unraveling Tissue-Specific Molecular Signatures and Convergent Pathway Enrichments in Suicidal Behavior

Suicide is a leading cause of death worldwide, yet the biological mechanisms underlying suicide remain poorly understood. A clearer understanding at the molecular level is essential for developing objective biomarkers and targeted interventions. In this study, we used transcriptomic profiling to investigate gene expression patterns associated with suicidal thoughts and behaviors across peripheral blood (n=264) and postmortem brain tissue from two prefrontal regions (dorsolateral prefrontal cortex, DLPFC; subgenual anterior cingulate cortex, sgACC) of individuals with and without psychiatric illness (n=249). Peripheral analyses revealed broad transcriptional changes associated with suicidal thoughts and behaviors, marked by dysregulated immune-related and inflammatory processes. Longitudinal modeling further revealed gene co-expression modules that predicted future suicide attempts over a 12-month follow-up, highlighting processes related to apoptosis, mitochondrial function, and immune regulation. By contrast, transcriptomic analyses of postmortem tissue derived from the DLPFC and sgACC revealed largely suppressed neuroimmune activity. Gene co-expression analyses in the brain identified suicide-associated modules enriched for synaptic plasticity, oxidative stress, and neuroimmune function, some of which displayed regional specificity. Cross-tissue comparison showed minimal gene-level overlap between brain and blood, although shared pathway-level themes emerged in immune, sensory, and cellular stress processes. Taken together, these findings suggest that suicide is associated with distinct but functionally convergent transcriptional alterations across brain and blood. By integrating tissue-specific and systems-level molecular signatures, this work provides insight into the biological architecture of suicide and lays the groundwork for developing novel biomarkers and therapeutic targets to improve prevention and treatment outcomes.

neuroscience↗

bioRxiv2026

Origin-1: a generative AI platform for de novo antibody design against novel epitopes

0Generative artificial intelligence has advanced antibody discovery, yet de novo design of therapeutic antibodies against targets with "zero-prior" epitopes remains a fundamental challenge. We define "zero-prior" epitopes as target sites lacking structural data from any reported antibody-antigen or protein-protein complex involving the target. Here we present Origin-1, a generative AI platform that overcomes this by integrating epitope-conditioned all-atom structure generation, paired complementarity determining region sequence design, and a specialized co-folding-based scoring protocol to select antibody designs predicted to be high-confidence, specific binders with favorable developability. We evaluated Origin-1 on a panel of ten targets selected to have no available protein-protein complex structures and minimal homology ([≤]60% sequence identity) to proteins with known complexes, creating stringent design conditions. In fewer than one hundred design attempts per target, we identified developable, specific antibodies, validated across multiple biophysical and developability assays, for four targets: COL6A3, AZGP1, CHI3L2, and IL36RA, with functional inhibition demonstrated for IL36RA. Cryogenic electron microscopy confirmed the atomic accuracy of our designs, revealing complexes that closely matched the computational models with high structural fidelity (3.0-3.3 [A] resolution; 0.83-0.91 DockQ). Furthermore, we employed AI-guided affinity maturation to optimize a de novo antibody binder against IL36RA, producing functional antagonists with sub-nanomolar affinities and a top EC50 of 12.3 nM. These results demonstrate a framework for targeting epitopes without structural precedent, expanding the programmable therapeutic antibody landscape.

molecular biology↗

bioRxiv2025

Permeabilization with fenchone enhances cryopreservation of Drosophila embryos

The difficulty of cryopreservation has long been a limitation of Drosophila melanogaster as a genetic model organism. Here we report a statistically significant improvement in the efficiency of Drosophila cryopreservation by substituting limonene with the monoterpenoid fenchone in the embryo permeabilization step of a previously published method. We found that fenchone-permeabilized embryos exhibit greater uptake of cryoprotectant compared with those permeabilized by limonene, and a ~6-fold increase in the rate of egg-to-adult survival for wild-type flies. Using this improved protocol, we successfully cryopreserved and revived precious strains after 12 months of storage in liquid nitrogen. These results suggest that fenchone is a superior permeabilizing agent for fly embryo cryopreservation, expanding possibilities for the long-term maintenance of Drosophila and other insect species. Further refinement of this approach may enable cryopreservation to replace continuous culture as the method of choice for routine maintenance of fly stocks.

physiology↗

bioRxiv2025

Dissecting Small Noncoding RNA Landscapes in Mouse Preimplantation Embryos and Human Blastoids for Modeling Early Human Embryogenesis

Small non-coding RNAs (sncRNA) exert regulatory functions in mammalian cells; however, their expression dynamics and contribution during preimplantation development are not entirely understood. In this study, we apply Small-Seq to generate a comprehensive single-cell atlas of sncRNA expression in mouse oocytes, sperm, and embryos (2-cell to 64-cell stages), and compare these dynamics with human embryos at equivalent stages. In both species, all sncRNA subtypes are expressed, but only microRNAs (miRNAs) and small nucleolar RNAs (snoRNAs) display cell type-specific patterns. In mice and humans, miRNAs and snoRNAs from the Dlk1-Dio3 locus were upregulated in the inner cell mass (ICM). Human trophectoderm (TE) is enriched with primate-specific C19MC miRNAs, which show relatively low expression in the ICM. In contrast, the mouse lacks a TE-specific hotspot, as C2MC miRNAs are higher in the ICM. Nonetheless, differentiation-associated miRNAs (e.g., miR-24-3p, miR-200c-3p) are high in the TE of both species. Further, we profile the sncRNAs of human blastoids (stem-cell-based embryo model) and determine that they largely recapitulate the human blastocyst. We envision broad utility of this dataset as a resource for future studies seeking to dissect the functions of individual sncRNAs in early development, and for advancing applications in stem cell-based embryo models and assisted reproductive technologies.

developmental biology↗

bioRxiv2025

GIN-CRC-Pareto: A graph-based Pareto-optimal multi-task learning framework to identify miRNA-target interactions in colorectal cancer

Colorectal cancer (CRC) ranks as the third highest incidence among malignancies in humans and the second most common cause of cancer-related mortality in the United States. Accumulating evidence has established microRNAs (miRNAs) as critical regulators of cancer development and therapeutic response. Understanding miRNA-mRNA interactions is critical for elucidating the molecular mechanisms driving CRC and other malignancies. In this study, we proposed GIN-CRC-Pareto, a graph-based, Pareto-optimal multi-task learning framework that simultaneously predicts miRNA-mRNA binding pairs, identifies seed match pairings, and classifies seed match subtypes. By leveraging the power of graph neural networks and Pareto-optimal gradient balancing strategy, GIN-CRC-Pareto dynamically adjusted the task weights during training to optimize each task without compromising the others. Experimental results demonstrated that our framework consistently outperforms traditional deep learning models and existing state-of-the-art tools across multiple evaluation metrics, with 0.909 in accuracy, 0.909 in precision and 0.968 in AUC in the miRNA-mRNA binding pairs prediction task. Additionally, we further validated the generalizability of the framework in combination with transfer learning techniques to identify miRNA-target interactions across other cancers. These findings highlight the effectiveness of the proposed framework to comprehensively identify the miRNA-target interactions in CRC, with the potential to serve as a scalable and generalizable tool across diverse cancer types, ultimately facilitating the development of miRNA-based therapeutics for cancer treatment.

bioinformatics↗

bioRxiv2025

Evaluation of Alphafold modeling for elucidation of nanobody-peptide epitope interactions

Models of Ab-antigen complexes can be used to understand interaction mechanisms and for improving affinity. This study evaluates the use of the protein structure prediction algorithm AlphaFold (AF) for exploration of interactions between peptide epitope tags and the smallest functional antibody fragments, nanobodies (Nbs). Although past studies of AF for modeling antibody-target (antigen) interactions suggested modest algorithm performance, those were primarily focused on Ab-protein interactions, while the performance and utility of AF for Nb-peptide interactions, which are generally less complex due to smaller antigens, smaller binding domains, and fewer chains, is less clear. In this study we evaluated the performance of AF for predicting the structures of Nbs bound to experimentally validated, linear, short peptide epitopes (Nb-tag pairs). We expanded the pool of experimental data available for comparison through crystallization and structural determination of a previously reported Nb-tag complex (Nb127). Models of Nb-tag pair structures generated from AF were variable with respect to consistency with experimental data, with good performance in just over half (4 out of 6) of cases. Even among Nb-tag pairs successfully modeled in isolation, efforts to translate modeling to more complex contexts failed, suggesting an underappreciated role of the size and complexity of inputs in AF modeling success. Finally, the model of a Nb-tag pair with minimal previous characterization was used to guide the design of a peptide-electrophile conjugate that undergoes covalent crosslinking with Nb upon binding. These findings highlight the utility of minimized antibody and antigen structures to maximize insights from AF modeling.

biochemistry↗

bioRxiv2025

SEHI-PPI: An End-to-End Sampling-Enhanced Human-Influenza Protein-Protein Interaction Prediction Framework with Double-View Learning

Influenza continues to pose significant global health threats, hijacking host cellular machinery through protein-protein interactions (PPIs), which are fundamental to viral entry, replication, immune evasion, and transmission. Yet, our understanding of these host-virus PPIs remains incomplete due to the vast diversity of viral proteins, their rapid mutation rates, and the limited availability of experimentally validated interaction data. Additionally, existing computational methods often struggle due to the limited availability of high-quality samples and their inability to effectively model the complex nature of host-virus interactions. To address these challenges, we present SEHI-PPI, an end-to-end framework for human-influenza PPI prediction. SEHI-PPI integrates a double-view deep learning architecture that captures both global and local sequence features, coupled with a novel adaptive negative sampling strategy to generate reliable and high-quality negative samples. Our method outperforms multiple benchmarks, including state-of-the-art large language models, achieving a superior performance in sensitivity (0.986) and AUROC (0.987). Notably, in a stringent test involving entirely unseen human and influenza protein families, SEHI-PPI maintains strong performance with an AUROC of 0.837. The model also demonstrates high generalizability across other human-virus PPI datasets, with an average sensitivity of 0.929 and AUROC of 0.928. Furthermore, AlphaFold3-guided case studies reveal that viral proteins predicted to target the same human protein cluster together structurally and functionally, underscoring the biological relevance of our predictions. These discoveries demonstrate the reliability of our SEHI-PPI framework in uncovering biologically meaningful host-virus interactions and potential therapeutic targets.

bioinformatics↗

bioRxiv2024

Blood-tumor barrier organoids recapitulate glioblastoma microenvironment and enable high-throughput modeling of therapeutic delivery

The blood-brain barrier (BBB) is a highly specialized system that is critical for regulating transport between the blood and the central nervous system. In brain tumors, the vasculature system is compromised, and is referred to as the blood-tumor barrier (BTB). The ability to precisely model the unique physiological properties of the BTB is essential to decipher its role in tumor pathophysiology and for the rational design of efficacious therapeutics. Here, we introduce a robust and high-throughput in vitro 3D human BTB organoid model that recapitulates various key features of the BTB observed in vivo and in clinical GBM samples. The organoids are composed of patient-derived glioblastoma stem cells (GSCs), human brain endothelial cells (EC), astrocytes and pericytes, which are formed through self-assembly. Transcriptomic and functional analyses reveal that the GSCs in the BTB organoids exhibit enhanced level of stemness, mesenchymal signature, invasiveness and angiogenesis, and this is further confirmed in in vivo studies. We demonstrate the ability of the BTB organoids to model therapeutic delivery and drug efficacy on brain tumor cells. Collectively, our findings show that the BTB organoid model has broad utility as a clinically representative system for studying the BTB and evaluating brain tumor therapies.

bioengineering↗

bioRxiv2024

A Comprehensive Engineering Strategy Improves Potency and Manufacturability of a Near Pan-neutralizing Antibody Against HIV

Anti-HIV envelope broadly neutralizing antibodies (bnAbs) are alternatives to conventional antiretrovirals with the potential to prevent and treat infection, reduce latent reservoirs, and/or mediate a functional cure. Clinical trials with "first generation" bnAbs used alone or in combination show promising antiviral effects but also highlight that additional engineering of "enhanced" antibodies will be required for optimal clinical utility, while preserving or enhancing cGMP manufacturing capability. Here we report the engineering of an anti-CD4 binding-site (CD4bs) bnAb, N49P9.3, purified from the plasma of an HIV elite-neutralizer. Through a series of rational modifications we produced a variant that demonstrates: enhanced potency; superior antiviral activity in combination with other bnAbs; low polyreactivity; and longer circulating half-life. Additional engineering for manufacturing produced a final variant, eN49P9, with properties conducive to cGMP production. Overall, these efforts demonstrate the feasibility of developing enhanced anti-CD4bs bnAbs with greatly improved antiviral properties as well as potential translational value.

bioengineering↗

bioRxiv2024

GenoM7GNet: An Efficient N7-methylguanosine Site Prediction Approach Based on a Nucleotide Language Model

N7-methylguanosine (m7G), one of the mainstream post-transcriptional RNA modifications, occupies an exceedingly significant place in medical treatments. However, classic approaches for identifying m7G sites are costly both in time and equipment. Meanwhile, the existing machine learning methods extract limited hidden information from RNA sequences, thus making it difficult to improve the accuracy. Therefore, we put forward to a deep learning network, called "GenoM7GNet," for m7G site identification. This model utilizes a Bidirectional Encoder Representation from Transformers (BERT) and is pretrained on nucleotide sequences data to capture hidden patterns from RNA sequences for m7G site prediction. Moreover, through detailed comparative experiments with various deep learning models, we discovered that the one-dimensional convolutional neural network (CNN) exhibits outstanding performance in sequence feature learning and classification. The proposed GenoM7GNet model achieved 0.953 in accuracy, 0.932 in sensitivity, 0.976 in specificity, 0.907 in Matthews Correlation Coefficient and 0.984 in Area Under the receiver operating characteristic Curve on performance evaluation. Extensive experimental results further prove that our GenoM7GNet model markedly surpasses other state-of-the-art models in predicting m7G sites, exhibiting high computing performance.

bioinformatics↗

bioRxiv2024

Molecular and cellular rhythms in excitatory and inhibitory neurons in the mouse prefrontal cortex

Previous studies have shown that there are rhythms in gene expression in the mouse prefrontal cortex (PFC); however, the contribution of different cell types and potential variation by sex has not yet been determined. Of particular interest are excitatory pyramidal cells and inhibitory parvalbumin (PV) interneurons, as interactions between these cell types are essential for regulating the excitation/inhibition balance and controlling many of the cognitive functions regulated by the PFC. In this study, we identify cell-type specific rhythms in the translatome of PV and pyramidal cells in the mouse PFC and assess diurnal rhythms in PV cell electrophysiological properties. We find that while core molecular clock genes are conserved and synchronized between cell types, pyramidal cells have nearly twice as many rhythmic transcripts as PV cells (35% vs. 18%). Rhythmic transcripts in pyramidal cells also show a high degree of overlap between sexes, both in terms of which transcripts are rhythmic and in the biological processes associated with them. Conversely, in PV cells, rhythmic transcripts from males and females are largely distinct. Moreover, we find sex-specific effects of phase on action potential properties in PV cells that are eliminated by environmental circadian disruption. Together, this study demonstrates that rhythms in gene expression and electrophysiological properties in the mouse PFC vary by both cell type and sex. Moreover, the biological processes associated with these rhythmic transcripts may provide insight into the unique functions of rhythms in these cells, as well as their selective vulnerabilities to circadian disruption. Significance statementThis is the first study to examine translatomic rhythms in the mouse PFC with cell-type specificity. We find that the core molecular clock cycles in phase across cell types, indicating that previously described daily oscillations in the cortical excitation/inhibition balance are not the consequence of a phase offset between PV and pyramidal cells. Nevertheless, rhythmic transcripts and their associated biological processes differ by both sex and cell type, suggesting that molecular rhythms may play a unique role in different cell types and between sexes. Therefore, our results, such as the enrichment of transcripts associated with mitochondrial function in PV cells from males, point towards possible cell and sex-specific mechanisms that could contribute to psychiatric and cognitive diseases upon rhythm disruption.

neuroscience↗

bioRxiv2024

Structural characterization and AlphaFold modeling of human T cell receptor recognition of NRAS cancer neoantigens

T cell receptors (TCRs) that recognize cancer neoantigens are important for anti-cancer immune responses and immunotherapy. Understanding the structural basis of TCR recognition of neoantigens provides insights into their exquisite specificity and can enable design of optimized TCRs. We determined crystal structures of a human TCR in complex with NRAS Q61K and Q61R neoantigen peptides and HLA-A1 MHC, revealing the molecular underpinnings for dual recognition and specificity versus wild-type NRAS peptide. We then used multiple versions of AlphaFold to model the corresponding complex structures, given the challenge of immune recognition for such methods. Interestingly, one implementation of AlphaFold2 (TCRmodel2) was able to generate accurate models of the complexes, while AlphaFold3 also showed strong performance, although success was lower for other complexes. This study provides insights into TCR recognition of a shared cancer neoantigen, as well as the utility and practical considerations for using AlphaFold to model TCR-peptide-MHC complexes.

immunology↗

bioRxiv2024

Gra-CRC-miRTar: The pre-trained nucleotide-to-graph neural networks to identify potential miRNA targets in colorectal cancer

Colorectal cancer (CRC) is the third most diagnosed cancer and the second deadliest cancer worldwide representing a major public health problem. In recent years, increasing evidence has shown that microRNA (miRNA) can control the expression of targeted human messenger RNA (mRNA) by reducing their abundance or translation, acting as oncogenes or tumor suppressors in various cancers, including CRC. Due to the significant up-regulation of oncogenic miRNAs in CRC, elucidating the underlying mechanism and identifying dysregulated miRNA targets may provide a basis for improving current therapeutic interventions. In this paper, we proposed Gra-CRC-miRTar, a pre-trained nucleotide-to-graph neural network framework, for identifying potential miRNA targets in CRC. Different from previous studies, we constructed two pre-trained models to encode RNA sequences and transformed them into de Bruijn graphs. We employed different graph neural networks to learn the latent representations. The embeddings generated from de Bruijn graphs were then fed into a Multilayer Perceptron (MLP) to perform the prediction tasks. Our extensive experiments show that Gra-CRC-miRTar achieves better performance than other deep learning algorithms and existing predictors. In addition, our analyses also successfully revealed 172 out of 201 functional interactions through experimentally validated miRNA-mRNA pairs in CRC. Collectively, our effort provides an accurate and efficient framework to identify potential miRNA targets in CRC, which can also be used to reveal miRNA target interactions in other malignancies, facilitating the development of novel therapeutics.

bioinformatics↗