Machine learning and deep learning to identifying subarachnoid haemorrhage macrophage-associated biomarkers by bulk and single-cell sequencing.

We investigated subarachnoid haemorrhage (SAH) macrophage subpopulations and identified relevant key genes for improving diagnostic and therapeutic strategies. SAH rat models were established, and brain tissue samples underwent single-cell transcriptome sequencing and bulk RNA-seq. Using single-cell data, distinct macrophage subpopulations, including a unique SAH subset, were identified. The hdWGCNA method revealed 160 key macrophage-related genes. Univariate analysis and lasso regression selected 10 genes for constructing a diagnostic model. Machine learning algorithms facilitated model development. Cellular infiltration was assessed using the MCPcounter algorithm, and a heatmap integrated cell abundance and gene expression. A 3 × 3 convolutional neural network created an additional diagnostic model, while molecular docking identified potential drugs. The diagnostic model based on the 10 selected genes achieved excellent performance, with an AUC of 1 in both training and validation datasets. The heatmap, combining cell abundance and gene expression, provided insights into SAH cellular composition. The convolutional neural network model exhibited a sensitivity and specificity of 1 in both datasets. Additionally, CD14, GPNMB, SPP1 and PRDX5 were specifically expressed in SAH-associated macrophages, highlighting its potential as a therapeutic target. Network pharmacology analysis identified some targeting drugs for SAH treatment. Our study characterised SAH macrophage subpopulations and identified key associated genes. We developed a robust diagnostic model and recognised CD14, GPNMB, SPP1 and PRDX5 as potential therapeutic targets. Further experiments and clinical investigations are needed to validate these findings and explore the clinical implications of targets in SAH treatment.

IPF-related new macrophage subpopulations and diagnostic biomarker identification - combine machine learning with single-cell analysis.

Idiopathic pulmonary fibrosis (IPF) is a chronic disease of unknown etiology that lacks a specific treatment. In IPF, macrophages play a key regulatory role as a major component of the lung immune system, especially during inflammation and fibrosis. However, our understanding of the cellular heterogeneity and molecular characterization of macrophages in IPF, as well as their relevance in the clinical setting, is relatively limited. In this study, we analyzed in-depth single-cell transcriptome sequencing (scRNA-seq) data from lung tissues of IPF patients, identified macrophage subpopulations in IPF, and probed their molecular characteristics and biological functions. hdWGCNA identified co-expressed gene modules of a subpopulation of IPF-associated macrophages (IPF-MΦ), and probed the IPF-MΦ by a machine-learning approach. hdWGCNA identified a subpopulation of IPF-associated macrophage subpopulations and probed the IPF-MΦ signature gene (IRMG) for its prognostic value, and a prediction model was developed on this basis. In addition, IPF-MΦ was obtained after recluster analysis of macrophages in IPF lung tissues. Coexpressed gene modules of IPF-MΦ were identified by hdWGCNA. Then, a machine learning approach was utilized to reveal the characteristic genes of IPF-MΦ, and a prediction model was built on this basis. In addition, we discovered a type of macrophage unique to IPF lung tissue named ATP5-MΦ. Its characteristic gene encodes a subunit of the mitochondrial ATP synthase complex, which is closely related to oxidative phosphorylation and proton transmembrane transport, suggesting that ATP5-MΦ may have higher ATP synthesis capacity in IPF lung tissue. This study provides new insights into the pathogenesis of IPF and provides a basis for evaluating disease prognosis and predictive medicine in IPF patients.

Single-cell dissection, hdWGCNA and deep learning reveal the role of oxidatively stressed plasma cells in ulcerative colitis.

Ulcerative colitis (UC) develops as a result of complex interactions between various cell types in the mucosal microenvironment. In this study, we aim to elucidate the pathogenesis of ulcerative colitis at the single-cell level and unveil its clinical significance. Using single-cell RNA sequencing and high-dimensional weighted gene co-expression network analysis, we identify a subpopulation of plasma cells (PCs) with significantly increased infiltration in UC colonic mucosa, characterized by pronounced oxidative stress. Combining 10 machine learning approaches, we find that the PC oxidative stress genes accurately distinguish diseased mucosa from normal mucosa (independent external testing AUC=0.991, sensitivity=0.986, specificity=0.909). Using MCPcounter and non-negative matrix factorization, we identify the association between PC oxidative stress genes and immune cell infiltration as well as patient heterogeneity. Spatial transcriptome data is used to verify the infiltration of oxidatively stressed PCs in colitis. Finally, we develop a gene-immune convolutional neural network deep learning model to diagnose UC mucosa in different cohorts (independent external testing AUC=0.984, sensitivity=95.9%, specificity=100%). Our work sheds light on the key pathogenic cell subpopulations in UC and is essential for the development of future clinical disease diagnostic tools.

Integrated multi-omics and artificial intelligence to explore new neutrophils clusters and potential biomarkers in sepsis with experimental validation.

Background: Sepsis, causing serious organ and tissue damage and even death, has not been fully elucidated. Therefore, understanding the key mechanisms underlying sepsis-associated immune responses would lead to more potential therapeutic strategies. Methods: Single-cell RNA data of 4 sepsis patients and 2 healthy controls in the GSE167363 data set were studied. The pseudotemporal trajectory analyzed neutrophil clusters under sepsis. Using the hdWGCNA method, key gene modules of neutrophils were explored. Multiple machine learning methods were used to screen and validate hub genes for neutrophils. SCENIC was then used to explore transcription factors regulating hub genes. Finally, quantitative reverse transcription-polymerase chain reaction was to validate mRNA expression of hub genes in peripheral blood neutrophils of two mice sepsis models. Results: We discovered two novel neutrophil subtypes with a significant increase under sepsis. These two neutrophil subtypes were enriched in the late state during neutrophils differentiation. The hdWGCNA analysis of neutrophils unveiled that 3 distinct modules (Turquoise, brown, and blue modules) were closely correlated with two neutrophil subtypes. 8 machine learning methods revealed 8 hub genes with high accuracy and robustness (ALPL, ACTB, CD177, GAPDH, SLC25A37, S100A8, S100A9, and STXBP2). The SCENIC analysis revealed that APLP, CD177, GAPDH, S100A9, and STXBP2 were significant associated with various transcriptional factors. Finally, ALPL, CD177, S100A8, S100A9, and STXBP2 significantly up regulated in peripheral blood neutrophils of CLP and LPS-induced sepsis mice models. Conclusions: Our research discovered new clusters of neutrophils in sepsis. These five hub genes provide novel biomarkers targeting neutrophils for the treatment of sepsis.

Single cells and TRUST4 reveal immunological features of the HFRS transcriptome.

The etiology of hemorrhagic fever with renal syndrome (HFRS) is significantly impacted by a variety of immune cells. Nevertheless, the existing techniques for sequencing peripheral blood T cell receptor (TCR) or B cell receptor (BCR) libraries in HFRS are constrained by both limitations and high costs. In this investigation, we utilized the computational tool TRUST4 to generate TCR and BCR libraries utilizing comprehensive RNA-seq data from peripheral blood specimens of HFRS patients. This facilitated the examination of clonality and diversity within immune libraries linked to the condition. Despite previous research on immune cell function, the underlying mechanisms remain intricate, and differential gene expression across immune cell types and cell-to-cell interactions within immune cell clusters have not been thoroughly explored. To address this gap, we performed clustering analysis on 11 cell subsets derived from raw single-cell RNA-seq data, elucidating characteristic changes in cell subset proportions under disease conditions. Additionally, we utilized CellChat, a tool for cell-cell communication analysis, to investigate the impact of MIF family, CD70 family, and GALECTIN family cytokines-known to be involved in cell communication-on immune cell subsets. Furthermore, hdWGCNA analysis identified core genes implicated in HFRS pathogenesis within T cells and B cells. Trajectory analysis revealed that most cell subsets were in a developmental stage, with high expression of transcription factors such as NFKB and JUN in Effector CD8+ T cells, as well as in Naive CD4+ T cells and Naive B cells. Our findings provide a comprehensive understanding of the dynamic changes in immune cells during HFRS pathogenesis, identifying specific V genes and J genes in TCR and BCR that contribute to advancing our knowledge of HFRS. These insights offer potential implications for the diagnosis and treatment of this autoimmune disease.

Single-cell analysis and machine learning identify psoriasis-associated CD8+ T cells serve as biomarker for psoriasis.

Psoriasis is a chronic inflammatory skin disease, the etiology of which has not been fully elucidated, in which CD8+ T cells play an important role in the pathogenesis of psoriasis. However, there is a lack of in-depth studies on the molecular characterization of different CD8+ T cell subtypes and their role in the pathogenesis of psoriasis. This study aims to further expound the pathogenesy of psoriasis at the single-cell level and to explore new ideas for clinical diagnosis and new therapeutic targets. Our study identified a unique subpopulation of CD8+ T cells highly infiltrated in psoriasis lesions. Subsequently, we analyzed the hub genes of the psoriasis-specific CD8+ T cell subpopulation using hdWGCNA and constructed a machine-learning prediction model, which demonstrated good efficacy. The model interpretation showed the influence of each independent variable in the model decision. Finally, we deployed the machine learning model to an online website to facilitate its clinical transformation.

hdWGCNA and Cellular Communication Identify Active NK Cell Subtypes in Alzheimer's Disease and Screen for Diagnostic Markers through Machine Learning.

BACKGROUND: Alzheimer's disease (AD) is a recognized complex and severe neurodegenerative disorder, presenting a significant challenge to global health. Its hallmark pathological features include the deposition of ß-amyloid plaques and the formation of neurofibrillary tangles. Given this context, it becomes imperative to develop an early and accurate biomarker model for AD diagnosis, employing machine learning and bioinformatics analysis. METHODS: In this study, single-cell data analysis was employed to identify cellular subtypes that exhibited significant differences between the diseased and control groups. Following the identification of NK cells, hdWGCNA analysis and cellular communication analysis were conducted to pinpoint NK cell subset with the most robust communication effects. Subsequently, three machine learning algorithms-LASSO, Random Forest, and SVM-RFE-were employed to jointly screen for NK cell subset modular genes highly associated with AD. A logistic regression diagnostic model was then designed based on these characterized genes. Additionally, a protein-protein interaction (PPI) networks of model genes was established. Furthermore, unsupervised cluster analysis was conducted to classify AD subtypes based on the model genes, followed by the analysis of immune infiltration in the different subtypes. Finally, Spearman correlation coefficient analysis was utilized to explore the correlation between model genes and immune cells, as well as inflammatory factors. RESULTS: We have successfully identified three genes (RPLP2, RPSA, and RPL18A) that exhibit a high association with AD. The nomogram based on these genes provides practical assistance in diagnosing and predicting patients' outcomes. The interconnected genes screened through PPI are intricately linked to ribosome metabolism and the COVID-19 pathway. Utilizing the expression of modular genes, unsupervised cluster analysis unveiled three distinct AD subtypes. Particularly noteworthy is subtype C3, characterized by high expression, which correlates with immune cell infiltration and elevated levels of inflammatory factors. Hence, it can be inferred that the establishment of an immune environment in AD patients is closely intertwined with the heightened expression of model genes. CONCLUSION: This study has not only established a valuable diagnostic model for AD patients but has also delved deeply into the pivotal role of model genes in shaping the immune environment of individuals with AD. These findings offer crucial insights into early AD diagnosis and patient management strategies.

Expression characteristics and potential function of non-coding RNA in mouse cortical cells.

Non-coding RNAs (ncRNAs) play essential regulatory functions in various physiological and pathological processes in the brain. To systematically characterize the ncRNA profile in cortical cells, we downloaded single-cell SMART-Seq v4 data of mouse cerebral cortex. Our results revealed that the ncRNAs alone are sufficient to define the identity of most cortical cell types. We identified 1,600 ncRNAs that exhibited cell type specificity, even yielding to distinguish microglia from perivascular macrophages with ncRNA. Moreover, we characterized cortical layer and region specific ncRNAs, in line with the results by spatial transcriptome (ST) data. By constructing a co-expression network of ncRNAs and protein-coding genes, we predicted the function of ncRNAs. By integrating with genome-wide association studies data, we established associations between cell type-specific ncRNAs and traits related to neurological disorders. Collectively, our study identified differentially expressed ncRNAs at multiple levels and provided the valuable resource to explore the functions and dysfunctions of ncRNAs in cortical cells.

Identification and validation of key biomarkers associated with macrophages in nonalcoholic fatty liver disease based on hdWGCNA and machine learning.

BACKGROUND: NAFLD has attracted increasing attention because of its high prevalence and risk of progression to cirrhosis or even hepatocellular carcinoma. Therefore, research into the root causes and molecular indicators of NAFLD is crucial. METHODS: We analyzed scRNA-seq data and RNA-seq data from normal and NAFLD liver samples. We utilized hdWGCNA to find module-related genes associated with the phenotype. Multiple machine learning algorithms were used to validate the model diagnostics and further screen for genes that are characteristic of NAFLD. The NAFLD mouse model was constructed using the MCD diet to validate the diagnostic effect of the genes. RESULTS: We identified a specific macrophage population called NASH-macrophages by single-cell sequencing analysis. Cell communication analysis and Pseudo-time trajectory analysis revealed the specific role and temporal distribution of NASH-macrophages in NAFLD. The hdWGCNA screening yielded 30 genes associated with NASH-macrophages, and machine learning algorithms screened and obtained two genes characterizing NAFLD. The immune infiltration indicated that these genes were highly associated with macrophages. Notably, we verified by RT-qPCR, IHC, and WB that MAFB and CX3CR1 are highly expressed in the MCD mouse model and may play important roles. CONCLUSIONS: Our study revealed a macrophage population that is closely associated with NAFLD. Using hdWGCNA analysis and multiple machine learning algorithms, we identified two NAFLD signature genes that are highly correlated with macrophages. Our findings may provide potential feature markers and therapeutic targets for NAFLD.

Machine learning and single cell RNA sequencing analysis identifies regeneration-related hepatocytes and highlights a Birc5-related model for identifying cell proliferative ability.

BACKGROUND: Partial hepatectomy (PHx) has been shown to induce rapid regeneration of adult liver under emergency conditions. Therefore, an in-depth investigation of the underlying mechanisms that govern liver regeneration following PHx is crucial for a comprehensive understanding of this process. METHOD: We analyzed scRNA-seq data from liver samples of normal and PHx-48-hour mice. Seven machine learning algorithms were utilized to screen and validate a gene signature that accurately identifies and predicts this population. Co-immunostaining of zonal markers with BIRC5 to investigate regional characteristics of hepatocytes post-PHx. RESULTS: Single cell sequencing results revealed a population of regeneration-related hepatocytes. Transcription factor analysis emphasized the importance of Hmgb1 transcription factor in liver regeneration. HdWGCNA and machine learning algorithm screened and obtained the key signature characterizing this population, including a total of 17 genes and the function enrichment analysis indicated their high correlation with cell cycle pathway. It is note-worthy that we inferred that Hmgb1 might be vital in the regeneration-related hepatocytes of PHx_48h group. Parallelly, Birc5 might be closely related to the regulation of liver regeneration, and positively correlated with Hmgb1. CONCLUSIONS: Our study has identified a distinct population of hepatocytes that are closely associated with liver regeneration. Through machine learning algorithms, we have identified a set of 17 genes that are highly indicative of the regenerative capacity of hepatocytes. This gene signature has enabled us to assess the proliferation ability of in vitro cultured hepatocytes using sequencing data alone.

Triosephosphate isomerase 1 may be a risk predictor in laryngeal squamous cell carcinoma: a multi-centered study integrating bulk RNA, single-cell RNA, and protein immunohistochemistry.

BACKGROUND: Although great progress has been made in anti-cancer therapy, the prognosis of laryngeal squamous cell carcinoma (LSCC) patients remains unsatisfied. Quantities of studies demonstrate that glycolytic reprograming is essential for the progression of cancers, where triosephosphate isomerase 1 (TPI1) serves as a catalytic enzyme. However, the clinicopathological significance and potential biological functions of TPI1 underlying LSCC remains obscure. METHODS: We collected in-house 82 LSCC tissue specimens and 56 non-tumor tissue specimens. Tissue microarrays (TMA) and immunohistochemical (IHC) experiments were performed. External LSCC microarrays and bulk RNA sequencing data were integrated to evaluate the expression of TPI1. We used a log-rank test and the CIBERSORT algorithm to assess the prognostic value of TPI1 and its association with the LSCC microenvironment. Malignant laryngeal epithelial cells and immune-stromal cells were identified using inferCNV and CellTypist. We conducted a comprehensive analysis to elucidate the molecular functions of TPI1 in LSCC tissue and single cells using Pearson correlation analysis, high dimensional weighted gene co-expression analysis, gene set enrichment analysis, and clustered regularly interspaced short palindromic repeats (CRISPR) screen. We explored intercellular communication patterns between LSCC single cells and immune-stromal cells and predicted several therapeutic agents targeting TPI1. RESULTS: Based on the in-house TMA and IHC analysis, TPI1 protein was found to have a strong positive expression in the nucleus of LSCC cells but only weakly positive activity in the cytoplasm of normal laryngeal cells (p < 0.0001). Further confirmation of elevated TPI1 mRNA expression was obtained from external datasets, comparing 251 LSCC tissue samples to 136 non-LSCC tissue samples (standardized mean difference = 1.06). The upregulated TPI1 mRNA demonstrated a high discriminative ability between LSCC and non-LSCC tissue (area under the curve = 0.91; sensitivity = 0.87; specificity = 0.79), suggesting its potential as a predictive marker for poor prognosis (p = 0.037). Lower infiltration abundance was found for plasma cells, naïve B cells, monocytes, and neutrophils in TPI-high expression LSCC tissue. Glycolysis and cell cycle were significantly enriched pathways for both LSCC tissue and single cells, where heat shock protein family B member 1, TPI1, and enolase 1 occupied a central position. Four outgoing communication patterns and two incoming communication patterns were identified from the intercellular communication networks. TPI1 was predicted as an oncogene in LSCC, with CRISPR scores less than -1 across 71.43% of the LSCC cell lines. TPI1 was positively correlated with the half maximal inhibitory concentration of gemcitabine and cladribine. CONCLUSIONS: TPI1 is dramatically overexpressed in LSCC than in normal tissue, and the high expression of TPI1 may promote LSCC deterioration through its metabolic and non-metabolic functions. This study contributes to advancing our knowledge of LSCC pathogenesis and may have implications for the development of targeted therapies in the future.

Identification of antigen-presentation related B cells as a key player in Crohn's disease using single-cell dissecting, hdWGCNA, and deep learning.

Crohn's disease (CD) arises from intricate intercellular interactions within the intestinal lamina propria. Our objective was to use single-cell RNA sequencing to investigate CD pathogenesis and explore its clinical significance. We identified a distinct subset of B cells, highly infiltrated in the CD lamina propria, that expressed genes related to antigen presentation. Using high-dimensional weighted gene co-expression network analysis and nine machine learning techniques, we demonstrated that the antigen-presenting CD-specific B cell signature effectively differentiated diseased mucosa from normal mucosa (Independent external testing AUC = 0.963). Additionally, using MCPcounter and non-negative matrix factorization, we established a relationship between the antigen-presenting CD-specific B cell signature and immune cell infiltration and patient heterogeneity. Finally, we developed a gene-immune convolutional neural network deep learning model that accurately diagnosed CD mucosa in diverse cohorts (Independent external testing AUC = 0.963). Our research has revealed a population of B cells with a potential promoting role in CD pathogenesis and represents a fundamental step in the development of future clinical diagnostic tools for the disease.