Predicted Binding Sites Research Articles

Discovery of Highly Potent Daphnane Diterpenoids Uncovers Importin-β1 as a Druggable Vulnerability in Castration-Resistant Prostate Cancer.

Importins are overexpressed in many cancers and mediate the abnormal nuclear transport of oncogenic factors. The druggable potential of importins still remains unclear, largely because of the lack of potent inhibitors. Herein, the anti-castration-resistant prostate cancer (CRPC) screening of a Euphorbiaceae diterpenoid library followed by target fishing led to the identification of a highly potent importin-β1 inhibitor, daphnane diterpenoid DD1. DD1 selectively inhibited the growth and survival of CRPC cells at subnanomolar concentrations and completely blocked tumor growth in preclinical models at an extremely low dosage. Mechanistic studies revealed that targeting of importin-β1 by DD1 significantly reduced the nuclear accumulation of key CRPC drivers, shutting down their downstream oncogenic signaling. Disruption of the predicted binding sites of DD1 on importin-β1 abolished this anti-CRPC effect. These findings suggest that importin-β1 is an effective therapeutic target in CRPC and that DD1 as the most potent importin-β1 inhibitor to date can be developed as therapeutics for treatment of this disease.

4D genetic networks reveal the genetic basis of metabolites and seed oil-related traits in 398 soybean RILs

BackgroundThe yield and quality of soybean oil are determined by seed oil-related traits, and metabolites/lipids act as bridges between genes and traits. Although there are many studies on the mode of inheritance of metabolites or traits, studies on multi-dimensional genetic network (MDGN) are limited.ResultsIn this study, six seed oil-related traits, 59 metabolites, and 107 lipids in 398 recombinant inbred lines, along with their candidate genes and miRNAs, were used to construct an MDGN in soybean. Around 175 quantitative trait loci (QTLs), 36 QTL-by-environment interactions, and 302 metabolic QTL clusters, 70 and 181 candidate genes, including 46 and 70 known homologs, were previously reported to be associated with the traits and metabolites, respectively. Gene regulatory networks were constructed using co-expression, protein–protein interaction, and transcription factor binding site and miRNA target predictions between candidate genes and 26 key miRNAs. Using modern statistical methods, 463 metabolite–lipid, 62 trait–metabolite, and 89 trait–lipid associations were found to be significant. Integrating these associations into the above networks, an MDGN was constructed, and 128 sub-networks were extracted. Among these sub-networks, the gene–trait or gene–metabolite relationships in 38 sub-networks were in agreement with previous studies, e.g., oleic acid (trait)–GmSEI–GmDGAT1a–triacylglycerol (16:0/18:2/18:3), gene and metabolite in each of 64 sub-networks were predicted to be in the same pathway, e.g., oleic acid (trait)–GmPHS–d-glucose, and others were new, e.g., triacylglycerol (16:0/18:1/18:2)–GmbZIP123–GmHD-ZIPIII-10–miR166s–oil content.ConclusionsThis study showed the advantages of MGDN in dissecting the genetic relationships between complex traits and metabolites. Using sub-networks in MGDN, 3D genetic sub-networks including pyruvate/threonine/citric acid revealed genetic relationships between carbohydrates, oil, and protein content, and 4D genetic sub-networks including PLDs revealed the relationships between oil-related traits and phospholipid metabolism likely influenced by the environment. This study will be helpful in soybean quality improvement and molecular biological research.

A reinforcement learning approach for protein–ligand binding pose prediction

Protein ligand docking is an indispensable tool for computational prediction of protein functions and screening drug candidates. Despite significant progress over the past two decades, it is still a challenging problem, characterized by the still limited understanding of the energetics between proteins and ligands, and the vast conformational space that has to be searched to find a satisfactory solution. In this project, we developed a novel reinforcement learning (RL) approach, the asynchronous advantage actor-critic model (A3C), to address the protein ligand docking problem. The overall framework consists of two models. During the search process, the agent takes an action selected by the actor model based on the current location. The critic model then evaluates this action and predict the distance between the current location and true binding site. Experimental results showed that in both single- and multi-atom cases, our model improves binding site prediction substantially compared to a naïve model. For the single-atom ligand, copper ion (Cu2+), the model predicted binding sites have a median root-mean-square-deviation (RMSD) of 2.39 Å to the true binding sites when starting from random starting locations. For the multi-atom ligand, sulfate ion (SO42−), the predicted binding sites have a median RMSD of 3.82 Å to the true binding sites. The ligand-specific models built in this study can be used in solvent mapping studies and the RL framework can be readily scaled up to larger and more diverse sets of ligands.

Transcription Factor Binding Site Prediction: Finding the Point from Many Data.

Transcription Factor Binding Site Prediction: Finding the Point from Many Data.

ProDFace: A web-tool for the dissection of protein-DNA interfaces.

Protein-DNA interactions play a crucial role in gene expression and regulation. Identifying the DNA binding surface of proteins has long been a challenge–in comparison to protein-protein interactions, limited progress has been made in the development of efficient DNA binding site prediction and protein-DNA docking methods. Here we present ProDFace, a web tool that characterizes the binding region of a protein-DNA complex based on amino acid propensity, hydrogen bond (HB) donor capacity (number of solvent accessible HB donor groups), sequence conservation at the interface core and rim region, and geometry. The program takes as input the structure of a protein-DNA complex in PDB (Protein Data Bank) format, and outputs various physicochemical and geometric parameters of the interface, as well as conservation of the interface residues in the protein component. Values are provided for the whole interface, and after dissecting it into core and rim regions. Details of water mediated HBs between protein and DNA, potential HB donor groups present at the binding surface of protein, and conserved interface residues are also provided as downloadable text files. These parameters can be useful in evaluating and validating protein-DNA docking solutions, structures derived from simulation as well as solutions from the available prediction tools, and facilitate the development of more efficient prediction methods. The web-tool is freely available at structbioinfo.iitj.ac.in/resources/bioinfo/pd_interface .

Flexibility-aware graph model for accurate epitope identification

Structural flexibility is an intrinsic characteristics of a protein upon interacting with other molecules, which mainly comes from the movement of a residue’s side chain, backbone and even an entire domain. Considering this property can be very helpful in protein binding analysis, such as epitope identification during antibody–antigen interaction. Unfortunately, to our knowledge, no approach is available at studying the dynamicity of protein binding from the computational perspective. We are pioneering a new perspective of exploring protein binding sites with considering the structural flexibility, particularly from the epitopes identification angle in antibody–antigen binding. To this end, we first obtained protein antigen structures with epitopes available, and built residue-level graphs of antigens. These graphs were highly densified subsequently by incorporating the structural flexibility. Later, the edge enriched graphs were clustered into overlapping subgraphs and were classified as epitope or non-epitope by a graph convolutional network. Experiments on epitope identification shown that the proposed flexibility-aware model markedly outperformed existing approaches by lifting the F1-score to 0.656, making a remarkable increment of 16.3% compared to the state-of-the-art. A quick study on generic protein binding site prediction also made a noteworthy improvement with increasing the F1-score by 8%. The superior performance obtained from both the specific and generic protein interaction analysis demonstrate that incorporating flexibility in computational models is helpful to strength the capability of identifying epitopes as well as general protein binding sites. This seminal study can be inspiring and promising to the wide range of protein interaction analysis.

In Silico Domain Structural Model Analysis of Coronavirus ORF1ab Polyprotein

The world today is battling with a coronavirus infection that is considered a global pandemic. Coronavirus infection is mainly attribute to the varying technique of the replication and release of different genomic components of the virus. The present study aims to establish the physical and chemical features, as well as the basic structural and functional properties of Coronavirus ORF1ab domain. A molecular approach was adopt in this study using the Swiss Model and Phyre2 server whereas the prediction of the active ligand binding sites was done using Phyre2. The analysis of the structure of the protein showed that it has good structural and heat stability, as well as better hydrophilic features and acidic in nature. Based on the Homology modeling, only two binding active sites were noted with catalytic function being mediated by Zn2+ as the metallic heterogeneous ligand for binding sites prediction. The proteins mostly exhibited helical secondary configurations. This study can help in predicting and understanding the role of this domain protein in active coronavirus infection.

MiR-702-5p ameliorates diabetic encephalopathy in db/db mice by regulating 12/15-LOX

The purpose of this study was to investigate the effect of miR-702-5p on diabetic encephalopathy (DE) and the interaction of miR-702-5p/12/15-LOX in the central nervous system (CNS). In this study, db/db mice were used as DE animal model and HT22 cells were treated with high-glucose (HG). Based on the bioinformatics prediction of possible binding sites between miR-702-5p and 12/15-LOX, we found that the expression of miR-702-5p was significantly down-regulated while 12/15-LOX up-regulated in vivo and in vitro, and the expression changes were inversely correlated. In vivo, diabetic mice with cognitive dysfunction and hippocampal neuronal damage had a concomitant increase in amyloid precursor protein (APP), amyloid beta(Aβ), tau, BAX protein expressions; by contrast, Bcl-2 protein expression was significantly decreased. Overexpression of miR-702-5p significantly reduced the histopathological damage of the hippocampus, improved the learning and memory function of db/db mice, down-regulated 12/15-LOX, APP, Aβ, tau, BAX protein expressions significantly and up-regulated the expression of Bcl-2. In vitro, miR-702-5p mimic reversed the decline in cell viability and the increase in cell apoptosis induced by HG. Simultaneously, reduced 12/15-LOX, APP, Aβ, BAX protein expressions, and increased Bcl-2 protein expression were detected in the miR-702-5p mimic group. Moreover, combined administration of miR-702-5p mimic and 12/15-LOX overexpression lentivirus significantly reversed the protective effect of up-regulation of miR-702-5p. In conclusion, miR-702-5p has a neuroprotective effect on DE, and this effect was achieved by inhibiting 12/15-LOX. However, miR-702-5p had an endogenous regulatory effect on 12/15-LOX rather than a direct targeting relationship.

NetTIME: a multitask and base-pair resolution framework for improved transcription factor binding site prediction.

MotivationMachine learning models for predicting cell-type-specific transcription factor (TF) binding sites have become increasingly more accurate thanks to the increased availability of next-generation sequencing data and more standardized model evaluation criteria. However, knowledge transfer from data-rich to data-limited TFs and cell types remains crucial for improving TF binding prediction models because available binding labels are highly skewed towards a small collection of TFs and cell types. Transfer prediction of TF binding sites can potentially benefit from a multitask learning approach; however, existing methods typically use shallow single-task models to generate low-resolution predictions. Here, we propose NetTIME, a multitask learning framework for predicting cell-type-specific TF binding sites with base-pair resolution.ResultsWe show that the multitask learning strategy for TF binding prediction is more efficient than the single-task approach due to the increased data availability. NetTIME trains high-dimensional embedding vectors to distinguish TF and cell-type identities. We show that this approach is critical for the success of the multitask learning strategy and allows our model to make accurate transfer predictions within and beyond the training panels of TFs and cell types. We additionally train a linear-chain conditional random field (CRF) to classify binding predictions and show that this CRF eliminates the need for setting a probability threshold and reduces classification noise. We compare our method’s predictive performance with two state-of-the-art methods, Catchitt and Leopard, and show that our method outperforms previous methods under both supervised and transfer learning settings.Availability and implementationNetTIME is freely available at https://github.com/ryi06/NetTIME and the code is also archived at https://doi.org/10.5281/zenodo.6994897.Supplementary information Supplementary data are available at Bioinformatics online.

IL12RB1 allele bias in human TH cells is regulated by functional SNPs in its 3′UTR

Allele bias is an epigenetic mechanism wherein only the maternal- or paternal-derived allele of a gene is preferentially expressed. Allele bias is used by T cells to regulate expression of numerous genes, including those which govern their development and response to cytokines. Here we demonstrate that human TH cell expression of the cytokine receptor gene IL12RB1 is subject to allele bias, and the extent to which this bias occurs is influenced by cells’ differentiation status and two polymorphic sites in the IL12RB1 3′UTR. The single nucleotide polymorphisms (SNPs) at these sites, rs3746190 and rs404733, function to increase expression of their encoding allele. Modeling suggests this is due to a stabilizing effect of these SNPs on the predicted mRNA secondary structure. The SNP rs3746190 is also proximal to the predicted binding site of microRNA miR-1277, raising the possibility that miR-1277 cannot exert suppression in the presence of rs3746190. Functional experiments demonstrate, however, that miR-1277 suppression of IL12RB1 3′UTR expression—which itself has not been previously reported—is nevertheless independent of rs3746190. Collectively, these data demonstrate that rs3746190 and rs404733 are functional SNPs which regulate IL12RB1 allele bias in human TH cells.

Transcriptome Analysis Reveals Gene Expression Changes during Repair from Mechanical Wounding in Aquilaria sinensis

Plants repair their mechanical wounds by reprogramming secondary metabolism. However, which genes are reprogrammed during this repair process in Aquilaria sinensis has rarely been studied. Here, we used high-throughput RNA sequencing to explore the changes in the transcriptome of Aquilaria’s xylem, six months after the stem was subjected to mechanical wounding. In total, 1165 transcripts were differentially accumulated, of which 1002 transcripts were increased and 163 were decreased in their abundances (|log2 (fold change)| ≥ 1 and FDR ≤ 0.05). The majority of these genes encode products involved in plant secondary metabolism, transcription regulation, and phytohormone metabolism and signaling. The up-regulated genes were classified into 15 significantly enriched GO terms and were involved in 83 pathways, whereas the down-regulated genes were classified into 5 significantly enriched GO terms and represented 43 pathways. Gene annotation demonstrated that 100 transcripts could encode transcription factors (TFs), such as WRKY, AP2, MYB, and Helix-loop-helix (HLH) TFs. We inferred that the differential expression of TFs, genes associated with plant hormones, phenylpropanoid biosynthesis, and sesquiterpenoid biosynthesis may contribute to the repair of the stem after mechanical wounding in A. sinensis. Using co-expression analysis and prediction of TF binding sites, a TF–gene regulatory network for Aquilaria lignin biosynthesis was constructed. This included the MYB, HLH, WRKY, and AP2 TFs, and the COMT1, 4CLL7, and CCR1 genes. The changes in 10 candidate genes were validated by quantitative reverse-transcription PCR, indicating significant differences between the treated and untreated areas. Our study provides global gene expression patterns under mechanical wounding and would be valuable to further studies on the molecular mechanisms of plant repair in A. sinensis.

Prediction of ligand binding sites using improved blind docking method with a Machine Learning-Based scoring function

The identification of the ligand binding sites (LBS) in proteins is of great significance for the elucidation of protein structure and function. Different methods have been published to predict protein–ligand binding sites efficiently. However, most of the current prediction methods only focus on the characteristics of proteins without considering the interactions between proteins and their different ligands, which often leads to frustrating results when the structure or function of the protein is complex or diverse. Therefore, an improved blind docking method with a machine learning-based scoring function is proposed in this paper for the LBS prediction. The blind docking method is used to search binding pockets and an artificial neural network is constructed to analyze binding features, which makes the proposed method possible to distinguish true binding sites from other possible pockets. Two cases of LBS prediction are presented to show the excellent performance of the proposed method. This paper aims to provide new ideas for the prediction of interactions between proteins and small molecules, which can further guide the research of structure-based drug discovery.

Interactions of Heavy Metals on Enzymes in Carbohydrate Metabolism: In silico study on glucokinase and pyruvate kinase

Cadmium (Cd2+) and mercury (Hg2+) are heavy metals, which can cause chronic inflammation, oxidative stress, obesity, hyperglycemia, and diabetes. Exposure to heavy metals cadmium and mercury can interfere with glycolysis metabolic functions through the inactivation of two key enzymes: glucokinase (GK) and pyruvate kinase (PyK). However, pathomechanism is unknown. Based on the background, this research was conducted. Enzyme structures were obtained from the RCSB Protein Data Bank (http://www.rcsb.org) with the following codes, GK (GDP ID: 3IDH) and PyK (GDP ID: 4IP7). The interaction between Cd and Hg with this enzyme was used via the MIB site: Metal Ion Binding site prediction and docking server (http://bioinfo.cmu.edu.tw/MIB/). Next, the target protein's interaction between metal ions and amino acids was visualized on UCSF Chimera 1.15. Based on the research results, the metal ion bond of mercury was more reactive than cadmium based on the number of amino acid residues bound. The bond was stronger based on a lower distance with PyK and GK enzymes. Therefore, mercury and cadmium metal ions were considered to inhibit the glycolysis process by causing the inactivation of the two enzymes.

MIB2: metal ion-binding site prediction and modeling server.

MIB2 (metal ion-binding) attempts to overcome the limitation of structure-based prediction approaches, with many proteins lacking a solved structure. MIB2 also offers more accurate prediction performance and more metal ion types. MIB2 utilizes both the (PS)2 method and the AlphaFold Protein Structure Database to acquire predicted structures to perform metal ion docking and predict binding residues. MIB2 offers marked improvements over MIB by collecting more MIB residue templates and using the metal ion type-specific scoring function. It offers a total of 18 types of metal ions for binding site predictions. Freely available on the web at http://bioinfo.cmu.edu.tw/MIB2/. Supplementary data are available at Bioinformatics online.

Prediction of protein mononucleotide binding sites using AlphaFold2 and machine learning

In this study, we developed a system that predicts the binding sites of proteins for five mononucleotides (AMP, ADP, ATP, GDP, and GTP). The system comprises two machine learning (ML)-based predictors using a convolutional neural network and a gradient boosting machine, two template-based predictors based on sequence and structure alignment, and a predictor that performs ensemble learning of these four predictors. In this study, data augmentation of ligand binding sites with similar ligand structures was performed. For example, in the prediction of ADP-binding sites using ML methods, the binding sites of AMP and ATP, which have similar structures, are considered. In addition, we constructed the structure models using AlphaFold2, a highly accurate protein prediction method. The secondary structure and dihedral angle information obtained using the model structures were used as ML predictor features. Additionally, in the template-based predictor, the structures of the binding sites were used as templates to be explored based on structure alignment to identify the binding site of the target. Consequently, the template-based predictor based on structure alignment showed the best performance among the four individual predictors, and the ensemble predictor achieved the best performance, with an area under the curve of 0.958 for all mononucleotides.

MiR-144 and DJ-1/NF-κB regulates UCP4 maintain mitochondrial homeostasis in Penaeus vannamei

UCP4, as an uncoupling protein in mitochondrial intima, is closely related to the resistance to oxidative stress and the function of mitochondria. However, whether and how its antioxidant capacity also works in crustaceans has not been reported in detail. This study showed that the expression of PvUCP4 was negatively correlated with the expression of pva-miR-144. The content of reactive oxygen species (ROS), ATP, and apoptosis was significantly increased, while the mitochondrial membrane potential (MMP) was seriously depolarized, Edema, vacuolation, and ambiguity of cristae and membrane were observed clearly in mitochondria after the knockdown of PvUCP4 induced by V. alginolyticus. The sharp drop in THC and severe damage in the hepatopancreas were all due to the knockout of PvUCP4 under the stress of V. alginolyticus. The co-transfection of pva-miR-144 and PvUCP4 could partially recover MMP compared with the abnormal expression of pva-miR-144. Similarly, co-transfection of pva-miR-144 and PvUCP4 could partially eliminate apoptosis compared with the abnormal expression of pva-miR-144. In addition, PvUCP4 3′-UTR has a pva-miR-144 predicted binding site in 1417–1428, which also was confirmed by the dual luciferase reporter assay. By the way, the results of ROS, MMP, and apoptosis showed that PvDJ-1 regulated the expression of PvUCP4 through PvNF-κB. Altogether, these results indicated that PvUCP4 has the antioxidant function of resisting oxidation reaction and weakening oxidative damage, to protect the normal operation of mitochondrial function and maintaining the cell homeostasis in shrimp.

Antisense Oligonucleotide Rescue of Deep-Intronic Variants Activating Pseudoexons in the 6-Pyruvoyl-Tetrahydropterin Synthase Gene.

We report two new 6-pyruvoyl-tetrahydropterin synthase splicing variants identified through genomic sequencing and transcript analysis in a patient with tetrahydrobiopterin deficiency, presenting with hyperphenylalaninemia and monoamine neurotransmitter deficiency. Variant c.243 + 3A>G causes exon 4 skipping. The deep-intronic c.164–672C>T variant creates a potential 5′ splice site that leads to the inclusion of four overlapping pseudoexons, corresponding to exonizations of an antisense short interspersed nuclear element AluSq repeat sequence. Two of the identified pseudoexons have been reported previously, activated by different deep-intronic variants, and were also detected at residual levels in control cells. Interestingly, the predominant pseudoexon is nearly identical to a disease causing activated pseudoexon in the F8 gene, with the same 3′ and 5′ splice sites. Splice switching antisense oligonucleotides (SSOs) were designed to hybridize with splice sites and/or predicted binding sites for regulatory splice factors. Different SSOs corrected the aberrant pseudoexon inclusion, both in minigenes and in fibroblasts from patients carrying the new variant c.164–672C>T or the previously described c.164–716A>T. With SSO treatment PTPS protein was recovered, illustrating the therapeutic potential of the approach, for patients with different pseudoexon activating variants in the region. In addition, the natural presence of pseudoexons in the wild type context suggests the possibility of applying the antisense strategy in patients with hypomorphic PTS variants with the purpose of upregulating their expression to increase overall protein and activity.

CavitySpace: A Database of Potential Ligand Binding Sites in the Human Proteome.

Location and properties of ligand binding sites provide important information to uncover protein functions and to direct structure-based drug design approaches. However, as binding site detection depends on the three-dimensional (3D) structural data of proteins, functional analysis based on protein ligand binding sites is formidable for proteins without structural information. Recent developments in protein structure prediction and the 3D structures built by AlphaFold provide an unprecedented opportunity for analyzing ligand binding sites in human proteins. Here, we constructed the CavitySpace database, the first pocket library for all the proteins in the human proteome, using a widely-applied ligand binding site detection program CAVITY. Our analysis showed that known ligand binding sites could be well recovered. We grouped the predicted binding sites according to their similarity which can be used in protein function prediction and drug repurposing studies. Novel binding sites in highly reliable predicted structure regions provide new opportunities for drug discovery. Our CavitySpace is freely available and provides a valuable tool for drug discovery and protein function studies.

ProB-Site: Protein Binding Site Prediction Using Local Features.

Protein–protein interactions (PPIs) are responsible for various essential biological processes. This information can help develop a new drug against diseases. Various experimental methods have been employed for this purpose; however, their application is limited by their cost and time consumption. Alternatively, computational methods are considered viable means to achieve this crucial task. Various techniques have been explored in the literature using the sequential information of amino acids in a protein sequence, including machine learning and deep learning techniques. The current efficiency of interaction-site prediction still has growth potential. Hence, a deep neural network-based model, ProB-site, is proposed. ProB-site utilizes sequential information of a protein to predict its binding sites. The proposed model uses evolutionary information and predicted structural information extracted from sequential information of proteins, generating three unique feature sets for every amino acid in a protein sequence. Then, these feature sets are fed to their respective sub-CNN architecture to acquire complex features. Finally, the acquired features are concatenated and classified using fully connected layers. This methodology performed better than state-of-the-art techniques because of the selection of the best features and contemplation of local information of each amino acid.

Identification of the hsa_circ_0039466/miR-96-5p/FOXO1 regulatory network in hepatocellular carcinoma by whole-transcriptome analysis.

BackgroundCircular RNAs (circRNAs) are important for the process of cancer initiation and progression. However, the role of circRNAs in hepatocellular carcinoma (HCC) remains incompletely understood. Therefore, we further explored the expression network of circRNAs in HCC.MethodsWhole-transcriptome microarrays of HCC and paired normal liver tissues were obtained from the Gene Expression Omnibus (GEO) database. The structures of tumor-associated circRNAs were acquired by the Cancer-Specific CircRNA Database (CSCD). StarBase, circBank, and R packages (miRNAtap and multiMiR) were used to predict miRNA targets of circRNAs and downstream molecules of miRNAs. Expression relationships between RNA-RNA interactions were evaluated by data from The Cancer Genome Atlas (TCGA) and GEO databases. ClusterProfiler and DOSE R packages were used for pathway enrichment to explore the biological functions of potential target genes. Finally, a possible circRNA-miRNA-mRNA regulatory network was established based on the competing endogenous RNA (ceRNA) hypothesis.ResultsThe differentially expressed circRNAs (DECs) were matched with cancer-specific circRNAs in the CSCD database and a screening analysis was performed to obtain 5 cancer-specific circRNAs. A total of 329 possible target miRNAs for 5 cancer-specific circRNAs were predicted by the circBank database, and intersection analysis with differentially expressed miRNAs (DEmiRNAs) revealed that miR-6746-3p and miR-96-5p were the two most suitable miRNAs targets for our selected circRNAs. Further expression verification and prediction of base complementary paired binding sites demonstrated the hsa_circ_0039466/miR-96-5p axis as a crucial pathway in HCC. Next, we found that FOXO1 and LEPR were two potential downstream molecules of the hsa_circ_0039466/miR-96-5p axis through target gene prediction analysis, differential expression analysis, and intersection analysis. The pathway enrichment results suggested that the hsa_circ_0039466/miR-96-5p axis affects the progression and outcome of HCC through the insulin resistance pathway. Finally, through multi-data crossover analysis and data analysis of HCC samples further confirmed the existence of the hsa_circ_0039466/miR-96-5p/FOXO1 ceRNA regulatory network and that the axis was closely related to clinical stage.Conclusionshsa_circ_0039466 facilitates the expression of FOXO1 by sponging miR-96-5p, and ultimately inhibits tumor progression. These results provide a theoretical basis for further understanding of the gene expression network of HCC.