ULDNA: integrating unsupervised multi-source language models with LSTM-attention network for high-accuracy protein-DNA binding site prediction.

Efficient and accurate recognition of protein-DNA interactions is vital for understanding the molecular mechanisms of related biological processes and further guiding drug discovery. Although the current experimental protocols are the most precise way to determine protein-DNA binding sites, they tend to be labor-intensive and time-consuming. There is an immediate need to design efficient computational approaches for predicting DNA-binding sites. Here, we proposed ULDNA, a new deep-learning model, to deduce DNA-binding sites from protein sequences. This model leverages an LSTM-attention architecture, embedded with three unsupervised language models that are pre-trained on large-scale sequences from multiple database sources. To prove its effectiveness, ULDNA was tested on 229 protein chains with experimental annotation of DNA-binding sites. Results from computational experiments revealed that ULDNA significantly improves the accuracy of DNA-binding site prediction in comparison with 17 state-of-the-art methods. In-depth data analyses showed that the major strength of ULDNA stems from employing three transformer language models. Specifically, these language models capture complementary feature embeddings with evolution diversity, in which the complex DNA-binding patterns are buried. Meanwhile, the specially crafted LSTM-attention network effectively decodes evolution diversity-based embeddings as DNA-binding results at the residue level. Our findings demonstrated a new pipeline for predicting DNA-binding sites on a large scale with high accuracy from protein sequence alone.

CircaKB: a comprehensive knowledgebase of circadian genes across multiple species.

Circadian rhythms, which are the natural cycles that dictate various physiological processes over a 24-h period, have been increasingly recognized as important in the management and treatment of various human diseases. However, the lack of sufficient data and reliable analysis methods have been a major obstacle to understanding the bidirectional interaction between circadian variation and human health. We have developed CircaKB, a comprehensive knowledgebase of circadian genes across multiple species. CircaKB is the first knowledgebase that provides systematic annotations of the oscillatory patterns of gene expression at a genome-wide level for 15 representative species. Currently, CircaKB contains 226 time-course transcriptome datasets, covering a wide variety of tissues, organs, and cell lines. In addition, CircaKB integrates 12 computational models to facilitate reliable data analysis and identify oscillatory patterns and their variations in gene expression. CircaKB also offers powerful functionalities to its users, including easy search, fast browsing, strong visualization, and custom upload. We believe that CircaKB will be a valuable tool and resource for the circadian research community, contributing to the identification of new targets for disease prevention and treatment. We have made CircaKB freely accessible at https://cdsic.njau.edu.cn/CircaKB.

HNSPPI: a hybrid computational model combing network and sequence information for predicting protein-protein interaction.

Most life activities in organisms are regulated through protein complexes, which are mainly controlled via Protein-Protein Interactions (PPIs). Discovering new interactions between proteins and revealing their biological functions are of great significance for understanding the molecular mechanisms of biological processes and identifying the potential targets in drug discovery. Current experimental methods only capture stable protein interactions, which lead to limited coverage. In addition, expensive cost and time consuming are also the obvious shortcomings. In recent years, various computational methods have been successfully developed for predicting PPIs based only on protein homology, primary sequences of protein or gene ontology information. Computational efficiency and data complexity are still the main bottlenecks for the algorithm generalization. In this study, we proposed a novel computational framework, HNSPPI, to predict PPIs. As a hybrid supervised learning model, HNSPPI comprehensively characterizes the intrinsic relationship between two proteins by integrating amino acid sequence information and connection properties of PPI network. The experimental results show that HNSPPI works very well on six benchmark datasets. Moreover, the comparison analysis proved that our model significantly outperforms other five existing algorithms. Finally, we used the HNSPPI model to explore the SARS-CoV-2-Human interaction system and found several potential regulations. In summary, HNSPPI is a promising model for predicting new protein interactions from known PPI data.

Systematically understanding the immunity leading to CRPC progression.

Prostate cancer (PCa) is the most commonly diagnosed malignancy and the second leading cause of cancer-related death in American men. Androgen deprivation therapy (ADT) has become a standard treatment strategy for advanced PCa. Although a majority of patients initially respond to ADT well, most of them will eventually develop castration-resistant PCa (CRPC). Previous studies suggest that ADT-induced changes in the immune microenvironment (mE) in PCa might be responsible for the failures of various therapies. However, the role of the immune system in CRPC development remains unclear. To systematically understand the immunity leading to CRPC progression and predict the optimal treatment strategy in silico, we developed a 3D Hybrid Multi-scale Model (HMSM), consisting of an ODE system and an agent-based model (ABM), to manipulate the tumor growth in a defined immune system. Based on our analysis, we revealed that the key factors (e.g. WNT5A, TRAIL, CSF1, etc.) mediated the activation of PC-Treg and PC-TAM interaction pathways, which induced the immunosuppression during CRPC progression. Our HMSM model also provided an optimal therapeutic strategy for improving the outcomes of PCa treatment.

Integrating MicroRNA Expression Profiling Studies to Systematically Evaluate the Diagnostic Value of MicroRNAs in Pancreatic Cancer and Validate Their Prognostic Significance with the Cancer Genome Atlas Data.

BACKGROUND/AIMS: MicroRNAs (miRNAs) are promising biomarkers for pancreatic cancer (PaCa). However, systemic and unified evaluations of the diagnostic value of miRNAs are lacking. Therefore, we performed a systematic evaluation based on miRNA expression profiling studies. METHODS: We obtained miRNA expression profiling studies from Gene Expression Omnibus (GEO) and ArrayExpress (AE) databases and calculated the pooled sensitivity, specificity, and summary area under a receiver operating characteristic (ROC) curve for every miRNA. According to the area under the curve (AUC), we identified the miRNAs with diagnostic potentiality and validated their prognostic role in The Cancer Genome Atlas (TCGA) data. Gene Ontology (GO) annotations and pathway enrichments of the target genes of the miRNAs were evaluated using bioinformatics tools. RESULTS: Ten miRNA expression profiling studies including 958 patients were used in this diagnostic meta-analysis. A total of 693 miRNAs were measured in more than 9 studies. The top 50 miRNAs with high predictive values for PaCa were identified. Among them, miR-130b had the best predictive value for PaCa (pooled sensitivity: 0.73 [95% confidence intervals (CI) 0.44-0.91], specificity: 0.81 [95% CI 0.59-0.93], and AUC: 0.84 [95% CI 0.73-0.95]). We identified nine miRNAs (miR-23a, miR-30a, miR-125a, miR-129-1, miR-181b-1, miR-203, miR-221, miR-222, and miR-1301) associated with overall survival in PaCa patients by combining our results with TCGA data. The results of a Cox model revealed that two miRNAs (miR-30a [hazard ratio (HR)=2.43, 95% CI 1.05-5.59; p=0.037] and miR-203 [HR=3.14, 95% CI 1.28-7.71; p=0.012]) were independent risk factors for prognosis in PaCa patients. In total, 405 target genes of the nine miRNAs were enriched with Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways, and cancer-associated pathways such as Ras signaling pathways, phospholipase D signaling pathway, and AMP-activated protein kinase (AMPK) signaling pathway were revealed among the top 20 enriched pathways. There were significant negative correlations between miR-181b-1 and miR-125a expression levels and the methylation status of their promoter region. CONCLUSION: Our study performed a systematic evaluation of the diagnostic value of miRNAs based on miRNA expression profiling studies. We identified that miR-23a, miR-30a, miR-125a, miR-129-1, miR-181b-1, miR-203, miR-221, miR-222, and miR-1301 had moderate diagnostic value for PaCa and predicted overall survival in PaCa patients.

Accelerating smooth molecular surface calculation.

This study proposes a novel approach, namely, skin flow complex algorithm (SFCA), to decompose the molecular skin surface into topological disks. The main contributions of SFCA include providing a simple decomposition and fast calculation of the molecular skin surface. Unlike most existing works which partition the molecular skin surface into sphere and hyperboloid patches, SFCA partitions the molecular skin surface into triangular quadratic patches and rectangular quadratic patches. Each quadratic patch is proven to be a topological disk and rendered by a rational Bézier patch. The skin surface is constructed by assembling all rational Bézier patches. Experimental results show that the SFCA is more efficient than most existing algorithms, and produces a triangulation of molecular skin surface which is decomposable, deformable, smooth, watertight and feature-preserved.

OS-SSVEP: One-shot SSVEP classification.

It is extremely challenging to classify steady-state visual evoked potentials (SSVEPs) in scenarios characterized by a huge scarcity of calibration data where only one calibration trial is available for each stimulus target. To address this challenge, we introduce a novel approach named OS-SSVEP, which combines a dual domain cross-subject fusion network (CSDuDoFN) with the task-related and task-discriminant component analysis (TRCA and TDCA) based on data augmentation. The CSDuDoFN framework is designed to comprehensively transfer information from source subjects, while TRCA and TDCA are employed to exploit the information from the single available calibration trial of the target subject. Specifically, CSDuDoFN uses multi-reference least-squares transformation (MLST) to map data from both the source subjects and the target subject into the domain of sine-cosine templates, thereby reducing cross-subject domain gap and benefiting transfer learning. In addition, CSDuDoFN is fed with both transformed and original data, with an adequate fusion of their features occurring at different network layers. To capitalize on the calibration trial of the target subject, OS-SSVEP utilizes source aliasing matrix estimation (SAME)-based data augmentation to incorporate into the training process of the ensemble TRCA (eTRCA) and TDCA models. Ultimately, the outputs of CSDuDoFN, eTRCA, and TDCA are combined for the SSVEP classification. The effectiveness of our proposed approach is comprehensively evaluated on three publicly available SSVEP datasets, achieving the best performance on two datasets and competitive performance on the third. Further, it is worth noting that our method follows a different technical route from the current state-of-the-art (SOTA) method and the two are complementary. The performance is significantly improved when our method is combined with the SOTA method. This study underscores the potential to integrate the SSVEP-based brain-computer interface (BCI) into daily life. The corresponding source code is accessible at https://github.com/Sungden/One-shot-SSVEP-classification.

Prediction of peptide drift time in ion mobility mass spectrometry from sequence-based features.

BACKGROUND: Ion mobility-mass spectrometry (IMMS), an analytical technique which combines the features of ion mobility spectrometry (IMS) and mass spectrometry (MS), can rapidly separates ions on a millisecond time-scale. IMMS becomes a powerful tool to analyzing complex mixtures, especially for the analysis of peptides in proteomics. The high-throughput nature of this technique provides a challenge for the identification of peptides in complex biological samples. As an important parameter, peptide drift time can be used for enhancing downstream data analysis in IMMS-based proteomics. RESULTS: In this paper, a model is presented based on least square support vectors regression (LS-SVR) method to predict peptide ion drift time in IMMS from the sequence-based features of peptide. Four descriptors were extracted from peptide sequence to represent peptide ions by a 34-component vector. The parameters of LS-SVR were selected by a grid searching strategy, and a 10-fold cross-validation approach was employed for the model training and testing. Our proposed method was tested on three datasets with different charge states. The high prediction performance achieve demonstrate the effectiveness and efficiency of the prediction model. CONCLUSIONS: Our proposed LS-SVR model can predict peptide drift time from sequence information in relative high prediction accuracy by a test on a dataset of 595 peptides. This work can enhance the confidence of protein identification by combining with current protein searching techniques.

scFseCluster: a feature selection-enhanced clustering for single-cell RNA-seq data.

Single-cell RNA sequencing (scRNA-seq) enables researchers to reveal previously unknown cell heterogeneity and functional diversity, which is impossible with bulk RNA sequencing. Clustering approaches are widely used for analyzing scRNA-seq data and identifying cell types and states. In the past few years, various advanced computational strategies emerged. However, the low generalization and high computational cost are the main bottlenecks of existing methods. In this study, we established a novel computational framework, scFseCluster, for scRNA-seq clustering analysis. scFseCluster incorporates a metaheuristic algorithm (Feature Selection based on Quantum Squirrel Search Algorithm) to extract the optimal gene set, which largely guarantees the performance of cell clustering. We conducted simulation experiments in several aspects to verify the performance of the proposed approach. scFseCluster performed very well on eight benchmark scRNA-seq datasets because of the optimal gene sets obtained using the Feature Selection based on Quantum Squirrel Search Algorithm. The comparative study demonstrated the significant advantages of scFseCluster over seven State-of-the-Art algorithms. In addition, our analysis shows that feature selection on high-variable genes can significantly improve clustering performance. In conclusion, our study demonstrates that scFseCluster is a highly versatile tool for enhancing scRNA-seq data clustering analysis.

Deep Transfer Learning for Cross-Species Plant Disease Diagnosis Adapting Mixed Subdomains.

A deep transfer learning framework adapting mixed subdomains is proposed for cross-species plant disease diagnosis. Most existing deep transfer learning studies focus on knowledge transfer between highly correlated domains. These methods may fail to deal with domains that are poorly correlated. In this study, mixed domain images were generated from source and target image groups for improving the correlation between the mixed domain (training dataset) and the target domain (testing dataset). A subdomain alignment mechanism is employed to transfer knowledge from the mixed domain to the target domain. The proposed framework captures the fine-grained information more effectively. Extensive experiments were conducted and prove that the proposed method produces a more effective result compared with existing deep transfer learning technologies for poorly related subdomains.

A Novel Feature Selection Strategy Based on Salp Swarm Algorithm for Plant Disease Detection.

Deep learning has been widely used for plant disease recognition in smart agriculture and has proven to be a powerful tool for image classification and pattern recognition. However, it has limited interpretability for deep features. With the transfer of expert knowledge, handcrafted features provide a new way for personalized diagnosis of plant diseases. However, irrelevant and redundant features lead to high dimensionality. In this study, we proposed a swarm intelligence algorithm for feature selection [salp swarm algorithm for feature selection (SSAFS)] in image-based plant disease detection. SSAFS is employed to determine the ideal combination of handcrafted features to maximize classification success while minimizing the number of features. To verify the effectiveness of the developed SSAFS algorithm, we conducted experimental studies using SSAFS and 5 metaheuristic algorithms. Several evaluation metrics were used to evaluate and analyze the performance of these methods on 4 datasets from the UCI machine learning repository and 6 plant phenomics datasets from PlantVillage. Experimental results and statistical analyses validated the outstanding performance of SSAFS compared to existing state-of-the-art algorithms, confirming the superiority of SSAFS in exploring the feature space and identifying the most valuable features for diseased plant image classification. This computational tool will allow us to explore an optimal combination of handcrafted features to improve plant disease recognition accuracy and processing time.

Automated CT Pancreas Segmentation for Acute Pancreatitis Patients by combining a Novel Object Detection Approach and U-Net.

Acute pancreatitis is an inflammatory disorder of the pancreas. Medical imaging, such as computed tomography (CT), has been widely used to detect volume changes in the pancreas for acute pancreatitis diagnosis. Many pancreas segmentation methods have been proposed but no methods for pancreas segmentation from acute pancreatitis patients. The segmentation of an inflamed pancreas is more challenging than the normal pancreas due to the following two reasons. 1) The inflamed pancreas invades surrounding organs and causes blurry boundaries. 2) The inflamed pancreas has higher shape, size, and location variability than the normal pancreas. To overcome these challenges, we propose an automated CT pancreas segmentation approach for acute pancreatitis patients by combining a novel object detection approach and U-Net. Our approach includes a detector and a segmenter. Specifically, we develop an FCN-guided region proposal network (RPN) detector to localize the pancreatitis regions. The detector first uses a fully convolutional network (FCN) to reduce the background interference of medical images and generates a fixed feature map containing the acute pancreatitis regions. Then the RPN is employed on the feature map to precisely localize the acute pancreatitis regions. After obtaining the location of pancreatitis, the U-Net segmenter is used on the cropped image according to the bounding box. The proposed approach is validated using a collected clinical dataset with 89 abdominal contrast-enhanced 3D CT scans from acute pancreatitis patients. Compared with other start-of-the-art approaches for normal pancreas segmentation, our method achieves better performance on both localization and segmentation in acute pancreatitis patients.

Evolution of the PB1 gene of human influenza A (H3N2) viruses circulating between 1968 and 2019.

One avian H3N2 influenza virus, providing its PB1 and HA segments, reassorted with one human H2N2 virus and caused a pandemic outbreak in 1968, killing over 1 million people. After its introduction to humanity, the pandemic H3N2 virus continued adapting to humans and has resulted in epidemic outbreaks every influenza season. To understand the functional roles of the originally avian PB1 gene in the circulating strains of human H3N2 influenza viruses, we analyzed the evolution of the PB1 gene in all human H3N2 isolates from 1968 to 2019. We found several specific residues dramatically changed around 2002-2009 and remained stable through to 2019. Then, we verified the functions of these PB1 mutations in the genetic background of the early pandemic virus, A/Hong Kong/1/1968(HK/68), as well as a recent seasonal strain, A/Jiangsu/34/2016 (JS/16). The PB1 V709I or PB1 V113A/K586R/D619N/V709I induced higher polymerase activity of HK/68 in human cells. And the four mutations acted cooperatively that had an increased replication capacity in vitro and in vivo at an early stage of infection. In contrast, the backward mutant, A113V/R586K/N619D/I709V, reduced polymerase activity in human cells. The PB1 I709V decreased viral replication in vitro, but this mutant only showed less effect on mice infection experiment, which suggested influenza A virus evolved in human host was not always consisted with highly replication efficiency and pathogenicity in other mammalian host. Overall, our results demonstrated that the identified PB1 mutations contributed to the viral evolution of human influenza A (H3N2) viruses.

Redox integration of signaling and metabolism in a head and neck cancer model of radiation resistance using COSMRO.

Redox metabolism is increasingly investigated in cancer as driving regulator of tumor progression, response to therapies and long-term patients' quality of life. Well-established cancer therapies, such as radiotherapy, either directly impact redox metabolism or have redox-dependent mechanisms of action defining their clinical efficacy. However, the ability to integrate redox information across signaling and metabolic networks to facilitate discovery and broader investigation of redox-regulated pathways in cancer remains a key unmet need limiting the advancement of new cancer therapies. To overcome this challenge, we developed a new constraint-based computational method (COSMro) and applied it to a Head and Neck Squamous Cell Cancer (HNSCC) model of radiation resistance. This novel integrative approach identified enhanced capacity for H2S production in radiation resistant cells and extracted a key relationship between intracellular redox state and cholesterol metabolism; experimental validation of this relationship highlights the importance of redox state in cellular metabolism and response to radiation.

ROR activation by Nobiletin enhances antitumor efficacy via suppression of IκB/NF-κB signaling in triple-negative breast cancer.

Triple-negative breast cancer (TNBC) is a heterogeneous disease characterized by poor response to standard therapies and therefore unfavorable clinical outcomes. Better understanding of TNBC and new therapeutic strategies are urgently needed. ROR nuclear receptors are multifunctional transcription factors with important roles in circadian pathways and other processes including immunity and tumorigenesis. Nobiletin (NOB) is a natural compound known to display anticancer effects, and our previous studies showed that NOB activates RORs to enhance circadian rhythms and promote physiological fitness in mice. Here, we identified several TNBC cell lines being sensitive to NOB, by itself or in combination. Cell and xenograft experiments showed that NOB significantly inhibited TNBC cell proliferation and motility in vitro and in vivo. ROR loss- and gain-of-function studies showed concordant effects of the NOB-ROR axis on MDA-MB-231 cell growth. Mechanistically, we found that NOB activates ROR binding to the ROR response elements (RRE) of the IκBα promoter, and NOB strongly inhibited p65 nuclear translocation. Consistent with transcriptomic analysis indicating cancer and NF-κB signaling as major pathways altered by NOB, p65-inducible expression abolished NOB effects, illustrating a requisite role of NF-κB suppression mediating the anti-TNBC effect of NOB. Finally, in vivo mouse xenograft studies showed that NOB enhanced the antitumor efficacy in mammary fat pad implanted TNBC, as a single agent or in combination with the chemotherapy agent Docetaxel. Together, our study highlights an anti-TNBC mechanism of ROR-NOB via suppression of NF-κB signaling, suggesting novel preventive and chemotherapeutic strategies against this devastating disease.

Multi-scale modeling for systematically understanding the key roles of microglia in AD development.

Alzheimer's disease (AD) is the leading cause of age-related dementia, affecting over 5 million people in the United States. Unfortunately, current therapies are largely palliative and several potential drug candidates have failed in late-stage clinical trials. Studies suggest that microglia-mediated neuroinflammation might be responsible for the failures of various therapies. Microglia contribute to Aß clearance in the early stage of neurodegeneration and may contribute to AD development at the late stage by releasing pro-inflammatory cytokines. However, the activation profile and phenotypic changes of microglia during the development of AD are poorly understood. To systematically understand the key role of microglia in AD progression and predict the optimal therapeutic strategy in silico, we developed a 3D multi-scale model of AD (MSMAD) by integrating multi-level experimental data, to manipulate the neurodegeneration in a simulated system. Based on our analysis, we revealed that how TREM2-related signal transduction leads to an imbalance in the activation of different microglia phenotypes, thereby promoting AD development. Our MSMAD model also provides an optimal therapeutic strategy for improving the outcome of AD treatment.

A Novel Computational Framework for Precision Diagnosis and Subtype Discovery of Plant With Lesion.

Plants are often attacked by various pathogens during their growth, which may cause environmental pollution, food shortages, or economic losses in a certain area. Integration of high throughput phenomics data and computer vision (CV) provides a great opportunity to realize plant disease diagnosis in the early stage and uncover the subtype or stage patterns in the disease progression. In this study, we proposed a novel computational framework for plant disease identification and subtype discovery through a deep-embedding image-clustering strategy, Weighted Distance Metric and the t-stochastic neighbor embedding algorithm (WDM-tSNE). To verify the effectiveness, we applied our method on four public datasets of images. The results demonstrated that the newly developed tool is capable of identifying the plant disease and further uncover the underlying subtypes associated with pathogenic resistance. In summary, the current framework provides great clustering performance for the root or leave images of diseased plants with pronounced disease spots or symptoms.

Inhibitory effect of selected hydrocolloids on 2-amino-1-methyl-6-phenylimidazo [4,5-b]pyridine (PhIP) formation in chemical models and beef patties.

2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) is a mutagen and a rodent carcinogen mainly formed in thermally processed muscle foods. Hydrocolloids are widely used as thickeners, gelling agents and stabilizers to improve food quality in the food industry. In this study, the inhibitory effects of eight hydrocolloids on the formation of PhIP were investigated in both chemical models and beef patties. 1% (w/w) of carboxymethylcellulose V, κ-carrageenan, alginic acid, and pectin significantly reduced PhIP formation by 53 %, 54 %, 48 %, and 47 %, respectively in chemical models. In fried beef patties, κ-carrageenan appeared to be most capable of inhibiting PhIP formation among the eight tested hydrocolloids. 1% (w/w) of κ-carrageenan caused a decreased formation of PhIP by 90 %. 1% (w/w) of κ-carrageenan also significantly reduced the formation of other heterocyclic aromatic amines including MeIQx and 4,8-DiMeIQx by 64 % and 48 %, respectively in fried beef patties. Further mechanism study showed that κ-carrageenan addition decreased the PhIP precursor creatinine residue and reduced the content of Maillard reaction intermediates including phenylacetaldehyde and aldol condensation product in the chemical model. κ-Carrageenan may inhibit PhIP formation via trapping both creatinine and phenylacetaldehyde. The structures of adducts formed between κ-carrageenan and creatinine and κ-carrageenan and phenylacetaldehyde merits further study.

Potential Pathogenic Genes Prioritization Based on Protein Domain Interaction Network Analysis.

Pathogenicity-related studies are of great importance in understanding the pathogenesis of complex diseases and improving the level of clinical medicine. This work proposed a bioinformatics scheme to analyze cancer-related gene mutations, and try to figure out potential genes associated with diseases from the protein domain-domain interaction network. Herein, five measures of the principle of centrality lethality had been adopted to implement potential correlation analysis, and prioritize the significance of genes. This method was further applied to KEGG pathway analysis by taking the malignant melanoma as an example. The experimental results show that 25 domains can be found, and 18 of them have high potential to be pathogenically important related to malignant melanoma. Finally, a web-based tool, named Human Cancer Related Domain Interaction Network Analyzer, is developed for potential pathogenic genes prioritization for 26 types of human cancers, and the analysis results can be visualized and downloaded online.

Single-cell transcriptomics reveals heterogeneous progression and EGFR activation in pancreatic adenosquamous carcinoma.

Pancreatic adenosquamous carcinoma (PASC) - a rare pathological pancreatic cancer (PC) type - has a poor prognosis due to high malignancy. To examine the heterogeneity of PASC, we performed single-cell RNA sequencing (scRNA-seq) profiling with sample tissues from a healthy donor pancreas, an intraductal papillary mucinous neoplasm, and a patient with PASC. Of 9,887 individual cells, ten cell subpopulations were identified, including myeloid, immune, ductal, fibroblast, acinar, stellate, endothelial, and cancer cells. Cancer cells were divided into five clusters. Notably, cluster 1 exhibited stem-like phenotypes expressing UBE2C, ASPM, and TOP2A. We found that S100A2 is a potential biomarker for cancer cells. LGALS1, NPM1, RACK1, and PERP were upregulated from ductal to cancer cells. Furthermore, the copy number variations in ductal and cancer cells were greater than in the reference cells. The expression of EREG, FCGR2A, CCL4L2, and CTSC increased in myeloid cells from the normal pancreas to PASC. The gene sets expressed by cancer-associated fibroblasts were enriched in the immunosuppressive pathways. We demonstrate that EGFR-associated ligand-receptor pairs are activated in ductal-stromal cell communications. Hence, this study revealed the heterogeneous variations of ductal and stromal cells, defined cancer-associated signaling pathways, and deciphered intercellular interactions following PASC progression.