Non-negative Matrix Factorization Clustering Research Articles

Identification of Aging and Young Subtypes for Predicting Colorectal Cancer Prognosis and Immunotherapy Responses.

Colorectal cancer (CRC) is critically related to aging and severely threatens human lives. To better explore the effects of aging on CRC progression and therapy outcome, a reliable aging subtypes identification of CRC is urgently desired. Here, 28 aging-related genes associated with the CRC prognosis were selected by univariate Cox analyses. Based on these 28 genes, CRC patients were divided into the aging subtype and young subtype by non-negative matrix factorization clustering. Aging subtype and young subtype of CRC were identified with distinct molecular features and clinical prognosis. The aging subtype was characterized by upregulation of senescence-associated secretory phenotype, higher frequencies of TP53 and immune checkpoint molecules, and high sensitivity to protein kinase and angiogenesis inhibitors. Furthermore, 14 genes were selected by LASSO penalized Cox regression analyses for aging-related risk signature construction. The constructed aging risk signature exhibited good prediction and the nomogram showed robust discrimination power over the traditional CRC staging system. In conclusion, this study successfully established aging subtype and young subtype of CRC, which is helpful to identify patients with aging characteristics to evaluate prognosis and treatment outcomes. Introducing aging-based subtypes into clinical concern and patient prognostication provides new opportunities for personalized CRC treatment.

Collaborative Learning based Recommendation System for Content Streaming Platform using Non-Negative Matrix Factorization Clustering

The Recommender System helps users find relevant content by leveraging their interest models. This paper focuses on content-based recommender systems, which determine recommendations based on previously selected products without needing user evaluations. These systems can now use machine learning to profile both consumers and items. We introduce a collaborative learning method for content streaming platforms using Non-Negative Matrix Factorization Clustering (NNMFC). The research employs a sliding window technique for clustering, implemented with gradient descent and further optimized using Particle Swarm Optimization (PSO). The model's performance was examined through its Root Mean Square Error (RMSE) across three techniques: gradient descent, sliding window gradient descent, and sliding window PSO. The result analysis of the proposed work achieved the lowest RMSE of 0.90, outperforming existing state-of-the-art approaches such as Collaborative Filtering, K-Mean, and UPCSim. The Sliding Window PSO emerged as the most proficient, presenting a notable 9% enhancement over existing methods.

NMF Clustering: Accessible NMF-based Clustering Utilizing GPU Acceleration.

Non-negative Matrix Factorization (NMF) is an algorithm that can reduce high dimensional datasets of tens of thousands of genes to a handful of metagenes which are biologically easier to interpret. Application of NMF on gene expression data has been limited by its computationally intensive nature, which hinders its use on large datasets such as single-cell RNA sequencing (scRNA-seq) count matrices. We have implemented NMF based clustering to run on high performance GPU compute nodes using CuPy, a GPU backed python library, and the Message Passing Interface (MPI). This reduces the computation time by up to three orders of magnitude and makes the NMF Clustering analysis of large RNA-Seq and scRNA-seq datasets practical. We have made the method freely available through the GenePattern gateway, which provides free public access to hundreds of tools for the analysis and visualization of multiple 'omic data types. Its web-based interface gives easy access to these tools and allows the creation of multi-step analysis pipelines on high performance computing (HPC) clusters that enable reproducible in silico research for non-programmers.

Disease patterns of coronary heart disease and type 2 diabetes harbored distinct and shared genetic architecture

Coronary heart disease (CHD) and type 2 diabetes (T2D) are two complex diseases with complex interrelationships. However, the genetic architecture of the two diseases is often studied independently by the individual single-nucleotide polymorphism (SNP) approach. Here, we presented a genotypic-phenotypic framework for deciphering the genetic architecture underlying the disease patterns of CHD and T2D. A data-driven SNP-set approach was performed in a genome-wide association study consisting of subpopulations with different disease patterns of CHD and T2D (comorbidity, CHD without T2D, T2D without CHD and all none). We applied nonsmooth nonnegative matrix factorization (nsNMF) clustering to generate SNP sets interacting the information of SNP and subject. Relationships between SNP sets and phenotype sets harboring different disease patterns were then assessed, and we further co-clustered the SNP sets into a genetic network to topologically elucidate the genetic architecture composed of SNP sets. We identified 23 non-identical SNP sets with significant association with CHD or T2D (SNP-set based association test, P < 3.70 × [Formula: see text]). Among them, disease patterns involving CHD and T2D were related to distinct SNP sets (Hypergeometric test, P < 2.17 × [Formula: see text]). Accordingly, numerous genes (e.g., KLKs, GRM8, SHANK2) and pathways (e.g., fatty acid metabolism) were diversely implicated in different subtypes and related pathophysiological processes. Finally, we showed that the genetic architecture for disease patterns of CHD and T2D was composed of disjoint genetic networks (heterogeneity), with common genes contributing to it (pleiotropy). The SNP-set approach deciphered the complexity of both genotype and phenotype as well as their complex relationships. Different disease patterns of CHD and T2D share distinct genetic architectures, for which lipid metabolism related to fibrosis may be an atherogenic pathway that is specifically activated by diabetes. Our findings provide new insights for exploring new biological pathways.

Integrated molecular analyses of an interferon-γ based subtype with regard to outcome, immune characteristics, and immunotherapy in bladder cancer and experimental verification

This study attempted to explore the role of interferon-γ related genes (IRGs) in the prognosis and immunotherapy of bladder cancer (BC). Based on data downloaded from public databases, molecular subtypes with different IRG expression patterns were determined via nonnegative matrix factorization clustering. On the basis of IRGs, interferon-γ related gene signature (IRGS) was developed through Cox regression analyses. We identified that two molecular subgroups with different outcome and immune profiles. It was proved that IRGS possessed prediction efficiency for BC prognosis. Compared with low IRGS group, high IRGS group was related to less anti-cancer immune cells infiltration, less tumor mutation burden score, more cancer stem cell index, and less benefit from immunotherapy. Differential expression of six model genes (IRF5, LATS2, MTHFD2, VAMP8, HLA-G and PTPN6) was validated between paired tissues by RT-qPCR. This study presents a prognostic model, which could serve as an indicator for the benefit of BC immunotherapy.

Dysregulation of HNF1B/Clusterin axis enhances disease progression in a highly aggressive subset of pancreatic cancer patients.

Pancreatic ductal adenocarcinoma (PDAC) is a lethal malignancy and is largely refractory to available treatments. Identifying key pathways associated with disease aggressiveness and therapeutic resistance may characterize candidate targets to improve patient outcomes. We used a strategy of examining the tumors from a subset of PDAC patient cohorts with the worst survival to understand the underlying mechanisms of aggressive disease progression and to identify candidate molecular targets with potential therapeutic significance. Non-negative matrix factorization (NMF) clustering, using gene expression profile, revealed three patient subsets. A 142-gene signature specific to the subset with the worst patient survival, predicted prognosis and stratified patients with significantly different survival in the test and validation cohorts. Gene-network and pathway analysis of the 142-gene signature revealed dysregulation of Clusterin (CLU) in the most aggressive patient subset in our patient cohort. Hepatocyte nuclear factor 1 b (HNF1B) positively regulated CLU, and a lower expression of HNF1B and CLU was associated with poor patient survival. Mechanistic and functional analyses revealed that CLU inhibits proliferation, 3D spheroid growth, invasiveness and epithelial-to-mesenchymal transition (EMT) in pancreatic cancer cell lines. Mechanistically, CLU enhanced proteasomal degradation of EMT-regulator, ZEB1. In addition, orthotopic transplant of CLU-expressing pancreatic cancer cells reduced tumor growth in mice. Furthermore, CLU enhanced sensitivity of pancreatic cancer cells representing aggressive patient subset, to the chemotherapeutic drug gemcitabine. Taken together, HNF1B/CLU axis negatively regulates pancreatic cancer progression and may potentially be useful in designing novel strategies to attenuate disease progression in PDAC patients.

A system for biomedical audio signal processing based on high performance computing techniques

In this paper, a noninvasive portable prototype is presented for biomedical audio signal processing. The proposed prototype is suitable for monitoring the health of patients. The proposed hardware setup consists of a cost-effective microphone, multipurpose microcontroller and computing node that could be a mobile phone or general-purpose computer. Using parallel and high-performance techniques, this setup allows one to register and wirelessly multicast the recorded biomedical signals to computing nodes in real time. The developed prototype was used as a case study to estimate the heart rate (HR) from the captured biomedical audio signal. In this regard, the developed algorithm for estimating HR comprises three stages: preprocessing, separation, and HR estimation. In the first stage, the signal captured by the microphone is adapted for processing. Subsequently, a separation stage was proposed to alleviate the acoustic interference between the lungs and heart. The separation is performed by combining a non-negative matrix factorization algorithm, clustering approach, and soft-filter strategy. Finally, HR estimation was obtained using a novel and efficient method based on the autocorrelation function. The developed prototype could be used not only for the estimation of the HR, but also for the retrieval of other biomedical information related to the recording of cardiac or respiratory audio signals. The proposed method was evaluated using well-known datasets and compared with state-of-the-art algorithms for source-separation. The results showed that it is possible to obtain an accurate separation and reliable real-time estimation in terms of source separation metrics and relative error in the tested scenarios by combining multi-core architectures with parallel and high-performance techniques. Finally, the proposed prototype was validated in a real-world scenario.

Identification of subtypes of hepatocellular carcinoma and screening of prognostic molecular diagnostic markers based on cell adhesion molecule related genes.

Cell adhesion molecules can predict liver hepatocellular carcinoma (LIHC) metastasis and determine prognosis, while the mechanism of the role of cell adhesion molecules in LIHC needs to be further explored. LIHC-related expression data were sourced from The Cancer Genome Atlas (TCGA) and the gene expression omnibus (GEO) databases, and genes related to cell adhesion were sourced from the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. First, the TCGA-LIHC dataset was clustered by the nonnegative matrix factorization (NMF) algorithm to find different subtypes of LIHC. Then the difference of prognosis and immune microenvironment between patients of different subtypes was evaluated. In addition, a prognostic risk model was obtained by least shrinkage and selection operator (LASSO) and Cox analysis, while a nomogram was drawn. Furthermore, functional enrichment analysis between high and low risk groups was conducted. Finally, the expressions of model genes were explored by quantitative real-time polymerase chain reaction (qRT-PCR). The 371 LIHC patients were classified into four subtypes by NMF clustering, and survival analysis revealed that disease-free survival (DFS) of these four subtypes were clearly different. Cancer-related pathways and immune microenvironment among these four subtypes were dysregulated. Moreover, 58 common differentially expressed genes (DEGs) between four subtypes were identified and were mainly associated with PPAR signaling pathway and amino acid metabolism. Furthermore, a prognostic model consisting of IGSF11, CD8A, ALCAM, CLDN6, JAM2, ITGB7, SDC3, CNTNAP1, and MPZ was built. A nomogram consisting of pathologic T and riskScore was built, and the calibration curve illustrated that the nomogram could better forecast LIHC prognosis. Gene Set Enrichment Analysis (GSEA) demonstrated that DEGs between high and low risk groups were mainly involved in cell cycle. Finally, the qRT-PCR illustrated the expressions of nine model genes between normal and LIHC tissue. A prognostic model consisting of IGSF11, CD8A, ALCAM, CLDN6, JAM2, ITGB7, SDC3, CNTNAP1, and MPZ was obtained, which provides an important reference for the molecular diagnosis of patient prognosis.

Metabolism-associated molecular classification of gastric adenocarcinoma.

Most gastric cancers (GC) are adenocarcinomas, whereas GC is a highly heterogeneous disease due to its molecular heterogeneity. However, traditional morphology-based classification systems, including the WHO classification and Lauren's classification, have limited utility in guiding clinical treatment. We performed nonnegative matrix factorization (NMF) clustering based on 2752 metabolism-associated genes. We characterized each of the subclasses from multiple angles, including subclass-associated metabolism signatures, immune cell infiltration, clinic10al characteristics, drug sensitivity, and pathway enrichment. As a result, four subtypes were identified: immune suppressed, metabolic, mesenchymal/immune exhausted and hypermutated. The subtypes exhibited significant prognostic differences, which suggests that the metabolism-related classification has clinical significance. Metabolic and hypermutated subtypes have better overall survival, and the hypermutated subtype is likely to be sensitive to anti-PD-1 immunotherapy. In addition, our work showed a strong connection with previously established classifications, especially Lei's subtype, to which we provided an interpretation based on the immune cell infiltration perspective, deepening the understanding of GC heterogeneity. Finally, a 120-gene classifier was generated to determine the GC classification, and a 10-gene prognostic model was developed for survival time prediction.

Evolutionary History of Transformation from Chronic Lymphocytic Leukemia to Richter Syndrome

Richter syndrome (RS), an aggressive lymphoma that develops in patients with chronic lymphocytic leukemia (CLL), is a striking example of histologic transformation. While recent advances have transformed the treatment landscape of CLL, RS remains associated with dismal overall survival and understanding of the genetic factors driving evolution of CLL to RS is limited. To decipher the genetics underlying this transformation, we performed an integrative analysis of whole exome (WES) and genome sequencing (WGS) data generated from matched RS and CLL samples from a discovery cohort of 52 patients with RS of diffuse large B cell lymphoma (DLBCL) histology followed by WES analysis of a validation cohort of 45 independent RS cases. Through computational deconvolution of CLL and RS clones, we constructed phylogenetic relationships and traced evolution of CLL to RS, confirming both clonal related (87%) and unrelated RS (13%). In addition to identifying recognized RS-risk genetic lesions, we discovered novel putative somatic driver genes (including IRF2BP2, SRSF1, B2M, DNMT3A, CCND3), copy number alterations (sCNAs) beyond CDKN2A/B loss and MYC gain (including del(15q15)[MGA], del(7q36.2)[EZH2, KMT2C, POT1], gain(1q23.1)[MCL1, NOTCH2] and arm level loss of 1p, 6q, 9p and 9q), whole genome duplication (WGD), and chromothripsis. Additional validation was obtained through re-analysis of 14 RS whole genomes (Klintman et al., Blood, 2021). Overall, our analysis highlighted alterations of NOTCH, DNA damage response and MAPK pathways as frequently preexisting in CLL. Alterations in pathways of MYC signaling, epigenetics, interferon/inflammatory signaling, cell cycle deregulation and immune evasion were newly occurring at transformation. To assess the degree of similarity between RS and DLBCL, non-negative matrix factorization (NMF) clustering was performed on the 97 RS WES samples and 304 previously characterized DLBCL samples (Chapuy et al., Nature Medicine, 2018) based on identified RS and known DLBCL drivers. The majority of RS (75 of 97) clustered separately from DLBCL, while 7 of 8 clonally unrelated RS cases clustered with DLBCL (P=6.75x10-6), with membership across the DLBCL molecular clusters. To further examine RS, NMF clustering was performed across the 97 RS samples, and 5 robust genomic subtypes were identified: RS-1: WGD; RS-2: Trisomy 12 with NOTCH1/SPEN mutations; RS-3: Mut-TP53 and Mut-NOTCH1; RS-4: Mut-EGR2 and Mut-SF3B1; and RS-5: TP53-altered RS without Mut-NOTCH1. Supervised and unsupervised transcriptome analysis of 36 of the 97 RS samples supported the identified genomic subtypes. The less-genome altered RS subtypes (RS-2, RS-4) had improved overall survival (3.3, 11.3, 5.0, 16.7 and 4.0 months for clonally related RS in subtypes 1-5, respectively; P = 0.0082). To investigate the stepwise evolution of CLL to RS at high resolution, single-cell RNA-sequencing was performed on diagnostic biopsy samples from 5 RS patients. We developed a novel tool, CNVSingle, to infer allele specific single-cell sCNAs that enabled identification of the single-cell clusters representing distinct CLL and RS genetic subclones as well as intermediate or transitional evolutionary states. Genetic changes often immediately preceded the transcriptional shifts towards RS and increasing chromothripsis was again observed as a hallmark of RS. RS cells displayed gene expression enriched in pathways of MYC activation and cell cycle, in line with similar analyses on paired RS and CLL bulk transcriptomes. Given the numerous RS-associated genomic features identified by our study, we assessed the feasibility of non-invasive detection of RS events in cell-free DNA (cfDNA) extracted from RS patient plasma by ultra-low pass-WGS sequencing of 46 plasma samples from 24 RS patients. Indeed, RS-specific sCNAs, chromothripsis, RS-specific driver mutations and whole genome doubling were detected in plasma cfDNA, distinct from circulating CLL. This included detection of RS in cfDNA of 6 of 8 patients (75%) at RS diagnosis and 2 of 7 (29%) patients within 10 months preceding RS diagnosis. In conclusion, we report molecular characterization of the largest RS cohort to date (111 patients), demonstrate that RS is genetically distinct from DLBCL and provide novel genetic and prognostic insights into this aggressive malignancy, with potential for clinical translation towards a more precise RS diagnosis.

DNA Methylation-Based Burkitt Lymphoma Epitypes Have Distinct Molecular and Clinical Features

Introduction: Genome-wide hypomethylation coupled with hypermethylation of tumor suppressor genes is a hallmark of many cancers. Consistent changes in DNA methylation patterns are also a feature of immune cells as they undergo terminal differentiation and maturation. In previous studies, it has been shown that malignant B cells harbor a combination of de novo methylation while retaining some epigenetic features characteristic of their cell of origin. Although information on the genetic features of Burkitt Lymphoma (BL) is expanding, limited information exists on the methylation landscape and its role in pathogenesis. We sought to identify consistent DNA methylation changes and their interplay between BL pathogenesis and other features such as EBV status and somatic mutations. We performed an integrative analysis using comprehensive genome-wide CpG methylation profiling, whole genome sequencing, and gene expression analysis. Methods: Samples were collected as part of the Burkitt Lymphoma Genome Sequencing Project (BLGSP). DNA from 218 primary BL tumors (154 pediatric (111 EBV+), 64 adult (19 EBV+)), 13 BL cell lines (6 EBV+), and 6 normal centroblast (CB) samples were profiled using the Illumina EPIC array. DNA and RNA from biopsies and cell lines underwent whole genome and RNA sequencing to enable the correlation of DNA methylation status with gene expression and genomic alterations. Nine of the BL samples were separately subjected to whole genome sequencing using the PromethION (Oxford Nanopore) to infer genome-wide methylation landscape. Publicly available data from 1595 B cell samples representing both normal subpopulations and tumor subtypes were used for further comparison. To search for methylation patterns within BL, beta values from the 10000 most variable CpGs were used as features for non-negative matrix factorization (NMF) clustering. Results: Clustering revealed two distinct epigenetic subgroups (epitypes) within BL, which we have named hyperBL (cluster1) and hypoBL (cluster2). We found pediatric and adult tumors to be proportionally split across the two epitypes; however, we noted an over-representation of EBV+ tumors in hyperBL. PCA analysis comparing BL with the additional B-cell samples revealed a separation between the epitypes, with hyperBL projecting farther along the trajectory of B cell differentiation, suggestive of a more terminally differentiated B-cell. Investigating the differential methylation between epitypes revealed hyperBL to be associated with greater hypermethylation compared to hypoBL and CBs. The hypermethylated CpGs in hyperBL mainly affected promoter regions within CpG islands while hypomethylated CpGs were mostly intergenic or in open seas. Using PromethION data to allow for genome-wide analyses, we found hypermethylated regions of hyperBL were enriched for binding motifs for several transcription factors involved in B cell development including: IRF4, PAX5, YY1, BCL6, SP1, STAT3, and KLF9. We used mitotic clock estimates based on DNA methylation (epiCMIT) and found hyperBL had significantly elevated values, suggesting a greater proliferative history of these tumors. To further establish a molecular basis for these epitypes, we searched for gene expression differences and distinct driver mutation patterns. The genomes of hyperBL tumors harbored an overall greater mutation burden and significantly higher mutation burden within regions affected by aberrant somatic hypermutation (aSHM), which was associated with a higher frequency of mutations attributed to the SBS9 mutational signature. Gene expression differences indicate hypoBL tumors have upregulation of IRF4 induced pathways. To determine the clinical relevance of the unique epigenetic clusters we conducted survival analyses and found hyperBL samples to have significantly inferior progression free survival. Conclusion: This work identifies novel epigenetic subgroups within BL with characteristic genetic differences between them. HyperBL is associated with hypermethylation, a higher mutation burden, elevated rates of aSHM, and inferior PFS. These results establish a subset of patients with distinct epigenetic, molecular and clinical features and further elucidates key mechanisms underlying BL pathogenesis. Together these data highlight the need and potential for strategies that exploit these epigenetic changes as a therapeutic target. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Molecular Classification of Genes Associated with Hypoxic Lipid Metabolism in Pancreatic Cancer.

Abnormal lipid metabolism often occurs under hypoxic microenvironment, which is an important energy supplement for cancer cell proliferation and metastasis. We aimed to explore the lipid metabolism characteristics and gene expression features of pancreatic ductal adenocarcinoma (PDAC) related to hypoxia and identify biomarkers for molecular classification based on hypoxic lipid metabolism that are evaluable for PDAC prognosis and therapy. The multiple datasets were analyzed integratively, including corresponding clinical information of samples. PDAC possesses a distinct metabolic profile and oxygen level compared with normal pancreatic tissues, according to the bioinformatics methods. In addition, a study on untargeted metabolomics using Ultra Performance Liquid Chromatography Tandem Mass Spectrometry(UPLC-MS) revealed lipid metabolites differences affected by oxygen. Analysis of PDAC gene expression profiling in The Cancer Genome Atlas (TCGA) revealed that the sphingolipid process correlates closely with HIF1α. According to the characters of HIF-1 and sphingolipid, samples can be clustered into three subgroups using non-negative matrix factorization clustering. In cluster2, patients had an increased survival time. Relatively high MUC16 mutation arises in cluster2 and may positively influence the cancer survival rates. This study explored the expression pattern of lipid metabolism under hypoxia microenvironment in PDAC. On the basis of metabolic signatures, we identified the prognosis subtypes linking lipid metabolism to hypoxia. The classifications may be conducive to developing personalized treatment programs targeting metabolic profiles.

Prognostic Model and Immune Infiltration of Ferroptosis Subcluster-Related Modular Genes in Gastric Cancer.

Background Gastric cancer (GC) is one of the gastrointestinal tumors with the highest mortality rate. The number of GC patients is still high. As a way of iron-dependent programmed cell death, ferroptosis activates lipid peroxidation and accumulates large reactive oxygen species. The role of ferroptosis in GC prognosis was underrepresented. The objective was to investigate the role of ferroptosis-related genes (FRGs) in the prognosis and development of GC. Methods Datasets of GC patients were obtained from the Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA) database that include clinical information and RNA seq data. Through nonnegative matrix factorization (NMF) clustering, we identified and unsupervised cluster analysis of the expression matrix of FRGs. And we constructed the co-expression network between genes and clinical characteristics by consensus weighted gene co-expression network analysis (WGCNA). The prognostic model was constructed by univariate and multivariate regression analysis. The potential mechanisms of development and prognosis in GC were explored by Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis, gene ontology (GO), tumor immune microenvironment (TIME), and tumor mutation burden (TMB). Results Two molecular subclusters with different expression patterns of FRGs were identified, which have significantly different survival states. Ferroptosis subcluster-related modular genes were identified by WGCNA. Based on 8 ferroptosis subcluster-related modular genes (collagen triple helix repeat containing 1 (CTHRC1), podoplanin (PDPN), procollagen-lysine,2-oxoglutarate 5-dioxygenase 2 (PLOD2), glutamine-fructose-6-phosphate transaminase 2 (GFPT2), ATP-binding cassette subfamily A member 1 (ABCA1), G protein-coupled receptor 176 (GPR176), serpin family E member 1 (SERPINE1), dual specificity phosphatase 1 (DUSP1)) and clinicopathological features, a nomogram was constructed and validated for their predictive efficiency on GC prognosis. Through receiver operating characteristic (ROC) analysis, the results showed that the area under the curve (AUC) of 1-, 3-, and 5-year survival were 0.721, 0.747, and 0.803, respectively, indicating that the risk-scoring model we constructed had good prognosis efficacy in GC. The degree of immune infiltration in high-risk group was largely higher than low-risk group. It indicated that the immune cells have a good response in high-risk group of GC. The TMB of high-risk group was higher, which could generate more mutations and was more conducive to the body's resistance to the development of cancer. Conclusion The risk-scoring model based on 8 ferroptosis subcluster-related modular genes has shown outstanding advantages in predicting patient prognosis. The interaction of ferroptosis in GC development may provide new insights into exploring molecular mechanisms and targeted therapies for GC patients.

Construction of an immune-related signature for predicting the ischemic events in patients undergoing carotid endarterectomy.

Background: Inflammatory responses have drawn more attention to atherosclerosis; however, the immune-related genes (IRGs) as a prognostic factor in atherosclerotic plaque remain to be fully elucidated. Here, the purpose of this study was to investigate whether the IRGs could be identified as a reliable biomarker for predicting ischemic events in patients undergoing carotid endarterectomy (CEA). Methods: Two datasets GSE97210 and GSE21545 were downloaded from the Gene Expression Omnibus (GEO) database. The dataset GSE97210 was used to explore the significant pathways and differentially expressed IRGs (DEIRGs) between plaques and controls, which were further screened to identify the prognostic DEIRGs in the GSE21545 dataset. The identification of molecular subgroups with the prognostic gene expression patterns was achieved through nonnegative matrix factorization (NMF) clustering. Functional analyses including GO, KEGG, GSVA, and GSEA analyses, and immune analyses including xCell and ssGSEA algorithms were conducted to elucidate the underlying mechanisms. The prognostic risk model was constructed using the LASSO algorithm and multivariate Cox regression analysis. Results: A total of 796 DEIRGs (including 588 upregulated and 208 downregulated) were identified. Nine prognostic DEIRGs were further screened with univariate Cox regression analysis. Two clusters with different prognosis were grouped based on the prognostic DEIRGs. Immune infiltration analysis shows that cluster 2 with a better prognosis presented with a higher immune response than cluster 1. A prognostic model based on seven IRGs (IL2RA, NR4A2, DES, ERAP2, SLPI, RASGRP1, and AGTR2) was developed and verified. Consistent with the immune analysis of the cluster, the immune infiltration in the low-risk group with a better prognosis was also more active than that in the high-risk group. Finally, a nomogram based on the seven genes was constructed, which might have future implications in clinical care. Conclusion: The expression of immune-related genes is correlated with the immune microenvironment of atherosclerotic patients and could be applied to predict the ischemic events in patients undergoing CEA accurately.

Two subtypes of cutaneous melanoma with distinct mutational signatures and clinico-genomic characteristics.

Background: To decipher mutational signatures and their associations with biological implications in cutaneous melanomas (CMs), including those with a low ultraviolet (UV) signature. Materials and Methods: We applied non-negative matrix factorization (NMF) and unsupervised clustering to the 96-class mutational context of The Cancer Genome Atlas (TCGA) cohort (N = 466) as well as other publicly available datasets (N = 527). To explore the feasibility of mutational signature-based classification using panel sequencing data, independent panel sequencing data were analyzed. Results: NMF decomposition of the TCGA cohort and other publicly available datasets consistently found two mutational signatures: UV (SBS7a/7b dominant) and non-UV (SBS1/5 dominant) signatures. Based on mutational signatures, TCGA CMs were classified into two clusters: UV-high and UV-low. CMs belonging to the UV-low cluster showed significantly worse overall survival and landmark survival at 1-year than those in the UV-high cluster; low or high UV signature remained the most significant prognostic factor in multivariate analysis. The UV-low cluster showed distinct genomic and functional characteristic patterns: low mutation counts, increased proportion of triple wild-type and KIT mutations, high burden of copy number alteration, expression of genes related to keratinocyte differentiation, and low activation of tumor immunity. We verified that UV-high and UV-low clusters can be distinguished by panel sequencing. Conclusion: Our study revealed two mutational signatures of CMs that divide CMs into two clusters with distinct clinico-genomic characteristics. Our results will be helpful for the clinical application of mutational signature-based classification of CMs.

Construction and validation model of necroptosis-related gene signature associates with immunity for osteosarcoma patients

Osteosarcoma is the most common malignant tumor in children and adolescents and its diagnosis and treatment still need to be improved. Necroptosis has been associated with many malignancies, but its significance in diagnosing and treating osteosarcoma remains unclear. The objective is to establish a predictive model of necroptosis-related genes (NRGs) in osteosarcoma for evaluating the tumor microenvironment and new targets for immunotherapy. In this study, we download the osteosarcoma data from the TARGET and GEO websites and the average muscle tissue data from GTEx. NRGs were screened by Cox regression analysis. We constructed a prediction model through nonnegative matrix factorization (NMF) clustering and the least absolute shrinkage and selection operator (LASSO) algorithm and verified it with a validation cohort. Kaplan–Meier survival time, ROC curve, tumor invasion microenvironment and CIBERSORT were assessed. In addition, we establish nomograms for clinical indicators and verify them by calibration evaluation. The underlying mechanism was explored through the functional enrichment analysis. Eight NRGs were screened for predictive model modeling. NRGs prediction model through NMF clustering and LASSO algorithm was established. The survival, ROC and tumor microenvironment scores showed significant statistical differences among subgroups (P < 0.05). The validation model further verifies it. By nomogram and calibration, we found that metastasis and risk score were independent risk factors for the poor prognosis of osteosarcoma. GO and KEGG analyses demonstrate that the genes of osteosarcoma cluster in inflammatory, apoptotic and necroptosis signaling pathways. The significant role of the correlation between necroptosis and immunity in promoting osteosarcoma may provide a novel insight into detecting molecular mechanisms and targeted therapy.

Comprehensive characterisation of immunogenic cell death in melanoma revealing the association with prognosis and tumor immune microenvironment.

Increasing evidence has highlighted the critical functions of immunogenic cell death (ICD) within many tumors. However, the therapeutic possibilities and mechanism of utilizing ICD in melanoma are still not well investigated. Melanoma samples involved in our study were acquired from The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO) databases. First, pan-cancer analysis of ICD systematically revealed its expression characteristics, prognostic values, mutation information, methylation level, pathway regulation relationship in multiple human cancers. The non-negative matrix factorization clustering was utilized to separate the TCGA-melanoma samples into two subtypes (i.e. C1 and C2) with different prognosis and immune microenvironment based on the expression traits of ICD. Then, LASSO-Cox regression analysis was utilized to determine an ICD-dependent risk signature (ICDRS) based on the differentially expressed genes (DEGs) between the two subtypes. Principal component analysis and t-distributed stochastic neighbor embedding analysis of ICDRS showed that high- and low-risk subpopulations could be clearly distinguished. Survival analysis and ROC curves in the training, internal validation, and external validation cohorts highlighted the accurate prognosis evaluation of ICDRS. The obvious discrepancies of immune microenvironment between the different risk populations might be responsible for the different prognoses of patients with melanoma. These findings revealed the close association of ICD with prognosis and tumor immune microenvironment. More importantly, ICDRS-based immunotherapy response and targeted drug prediction might be beneficial to different risk subpopulations of patients with melanoma. The innotative ICDRS could function as a marker to determine the prognosis and tumor immune microenvironment in melanoma. This will aid in patient classification for individualized melanoma treatment.

Pyroptosis: a novel signature to predict prognosis and immunotherapy response in gliomas.

Gliomas are the most common primary brain tumors and are highly malignant with a poor prognosis. Pyroptosis, an inflammatory form of programmed cell death, promotes the inflammatory cell death of cancer. Studies have demonstrated that pyroptosis can promote the inflammatory cell death (ICD) of cancer, thus affecting the prognosis of cancer patients. Therefore, genes that control pyroptosis could be a promising candidate bio-indicator in tumor therapy. The function of pyroptosis-related genes (PRGs) in gliomas was investigated based on the Chinese Glioma Genome Atlas (CGGA), the Cancer Genome Atlas (TCGA) and the Repository of Molecular Brain Neoplasia Data (Rembrandt) databases. In this study, using the non-negative matrix factorization (NMF) clustering method, 26 PRGs from the RNA sequencing data were divided into two subgroups. The LASSO and Cox regression was used to develop a 4-gene (BAX, Caspase-4, Caspase-8, PLCG1) risk signature, and all glioma patients in the CGGA, TCGA and Rembrandt cohorts were divided into low- and high-risk groups. The results demonstrate that the gene risk signature related to clinical features can be used as an independent prognostic indicator in glioma patients. Moreover, the high-risk subtype had rich immune infiltration and high expression of immune checkpoint genes in the tumor immune microenvironment (TIME). The analysis of the Submap algorithm shows that patients in the high-risk group could benefit more from anti-PD1 treatment. The risk characteristics associated with pyroptosis proposed in this study play an essential role in TIME and can potentially predict the prognosis and immunotherapeutic response of glioma patients.

Lipid metabolism patterns and relevant clinical and molecular features of coronary artery disease patients: an integrated bioinformatic analysis

BackgroundHyperlipidaemia is an important factor that induces coronary artery disease (CAD). This study aimed to explore the lipid metabolism patterns and relevant clinical and molecular features of coronary artery disease patients.MethodsIn the current study, datasets were fetched from the Gene Expression Omnibus (GEO) database and nonnegative matrix factorization clustering was used to establish a new CAD classification based on the gene expression profile of lipid metabolism genes. In addition, this study carried out bioinformatics analysis to explore intrinsic biological and clinical characteristics of the subgroups.ResultsData for a total of 615 samples were extracted from the Gene Expression Omnibus database and were associated with clinical information. Then, this study used nonnegative matrix factorization clustering for RNA sequencing data of 581 lipid metabolism relevant genes, and the 296 patients with CAD were classified into three subgroups (NMF1, NMF2, and NMF3). Subjects in subgroup NMF2 tended to have an increased severity of CAD. The CAD index and age of group NMF1 were similar to those of group NMF3, but their intrinsic biological characteristics exhibited significant differences. In addition, weighted gene coexpression network analysis (WGCNA) was used to determine the most important modules and screen lipid metabolism related genes, followed by further analysis of the DEGs in which the significant genes were identified based on clinical information. The progression of coronary atherosclerosis may be influenced by genes such as PTGDS and DGKE.ConclusionDifferent CAD subgroups have their own intrinsic biological characteristics, indicating that more personalized treatment should be provided to patients in each subgroup, and some lipid metabolism related genes (PDGTS, DGKE and so on) were related significantly with clinical characteristics.

Identification of novel lactate metabolism signatures and molecular subtypes for prognosis in hepatocellular carcinoma.

Background: Evidence has shown that lactate, an immune signaling molecule, is associated with hepatocellular carcinoma (HCC) progression and immune suppression. Therefore, identifying lactate metabolism-related molecules is a promising therapeutic strategy to inhibit the development of HCC and overcome chemotherapy resistance. Long noncoding RNAs (lncRNAs) are related to tumorigenesis and metastasis. Hence, verifying the molecular subtypes of lncRNAs related to lactate metabolism will play a critical role in managing HCC. Methods: Based on HCC data in The Cancer Genome Atlas (TCGA), lactate metabolic pathway-related genes were enriched by gene collection and enrichment analysis (GSEA). Lactate metabolism-related lncRNAs (LM_lncRNAs) were identified by correlation analysis, HCC molecular subtypes were determined using nonnegative matrix factorization (NMF) clustering, and the response of the three subtypes to chemotherapeutics was further evaluated using the Genomic Tumor Sensitive Cell Line (GDSC) dataset. LM_lncRNAs were examined via Lasso-Cox regression analysis to determine prognosis for patients. A Nomagram plot was used to predict patient survival time. Results: Three molecular subtypes of HCC were identified. The survival rate of patients with C1 subtype was higher than that of those with C2 and C3. Additionally, patients with C3 subtype have higher levels of immune cell infiltration and high expression of genes related to immune checkpoints. The GDSC results indicated that patients with C3 subtypes were more sensitive to chemotherapy drugs such as sorafenib and sunitinib. The prognostic risk assessment model consisted of six risk factors (AC034229.4, AC131009.1, MYOSLID, AC008667.1, AC012073.1, AC068025.1) and two protective factors (LINC00402 and AC103858.1). Based on Kaplan-Meier analysis, low-risk HCC patients had a high survival rate, and the receiver operating characteristic curve (ROC), calibration curve, and C-index confirmed good prediction ability. Conclusion: In this study, the molecular subtyping method and prediction model of lactate metabolism-related lncRNAs (LM_lncRNAs) were constructed for the prognosis of HCC patients. This work demonstrated the potential targets of LM_lncRNAs and provided a novel perspective and therapeutic paradigm for future clinical translation.