Representative Gene Sets Research Articles

Comprehensive analysis of the interaction of antigen presentation during anti-tumour immunity and establishment of AIDPS systems in ovarian cancer.

There are hundreds of prognostic models for ovarian cancer. These genes are based on different gene classes, and there are many ways to construct the models. Therefore, this paper aims to build the most stable prognostic evaluation system known to date through 101 machine learning strategies. We combined 101 algorithm combinations with 10 machine learning algorithms to create antigen presentation-associated genetic markers (AIDPS) with outstanding precision and steady performance. The inclusive set of algorithms comprises the elastic network (Enet), Ridge, stepwise Cox, Lasso, generalized enhanced regression model (GBM), random survival forest (RSF), supervised principal component (SuperPC), Cox partial least squares regression (plsRcox), survival support vector machine (Survival-SVM). Then, in the train cohort, the prediction model was fitted using a leave-one cross-validation (LOOCV) technique, which involved 101 different possible combinations of prognostic genes. Seven validation data sets (GSE26193, GSE26712, GSE30161, GSE63885, GSE9891, GSE140082 and ICGC_OV_AU) were compared and analysed, and the C-index was calculated. Finally, we collected 32 published ovarian cancer prognostic models (including mRNA and lncRNA). All data sets and prognostic models were subjected to a univariate Cox regression analysis, and the C-index was calculated to demonstrate that the antigen presentation process should be the core criterion for evaluating ovarian cancer prognosis. In a univariate Cox regression analysis, 22 prognostic genes were identified based on the expression profiles of 283 genes involved in antigen presentation and the intersection of genes (p < 0.05). AIDPS were developed by our machine learning-based integration method, which was applied to these 22 genes. One hundred and one prediction models are fitted using the LOOCV framework, and the C-index is calculated for each model across all validation sets. Interestingly, RSF + Lasso was the best model overall since it had the greatest average C-index and the highest C-index of any combination of models tested on the validated data sets. In comparing external cohorts, we found that the C-index correlated AIDPS method using the RSF + Lasso method in 101 prediction models was in contrast to other features. Notably, AIDPS outperformed the vast majority of models across all data sets. Antigen-presenting anti-tumour immune pathways can be used as a representative gene set of ovarian cancer to track the prognosis of patients with cancer. The antigen-presenting model obtained by the RSF + Lasso method has the best C-INDEX, which plays a key role in developing antigen-presenting targeted drugs in ovarian cancer and improving the treatment outcome of patients.

MCM2 is involved in subtyping and tamoxifen resistance of ERα-positive breast cancer by acting as the downstream factor of ERα.

Tamoxifen (TAM) resistance is finally developed in over 40% of patients with estrogen receptor α-positive breast cancer (ERα+ -BC), documenting that discovering new molecular subtype is needed to confer perception to the heterogeneity of ERα+ -BC. We obtained representative gene sets subtyping ERα+ -BC using gene set variation analysis (GSVA), non-negative matrix factorization (NMF), and COX regression methods on the basis of METABRIC, TCGA, and GEO databases. Furthermore, the risk score of ERα+ -BC subtyping was established using least absolute shrinkage and selection operator (LASSO) regression on the basis of genes in the representative gene sets, thereby generating the two subtypes of ERα+ -BC. We further found that minichromosome maintenance complex component 2 (MCM2) functioned as the hub gene subtyping ERα+ -BC using GO, KEGG, and MCODE. MCM2 expression was capable for specifically predicting 1-year overall survival (OS) of ERα+ -BC and correlated with T stage, AJCC stage, and tamoxifen (TAM) sensitivity of ERα+ -BC. The downregulation of MCM2 expression inhibited proliferation, migration, and invasion of TAM-resistant cells and promoted G0/G1 arrest. Altogether, tamoxifen resistance entails that MCM2 is a hub gene subtyping ERα+ -BC, providing a novel dimension for discovering a potential target of TAM-resistant BC.

Role of estrogen receptor signaling pathway-related genes in diffuse large B-cell lymphoma and identification of key targets via integrated bioinformatics analysis and experimental validation.

Diffuse large B-cell lymphoma (DLBCL) is a highly heterogeneous malignancy. Epidemiologically, the incidence of DLBCL is higher in men, and the female sex is a favorable prognostic factor, which can be explained by estrogen. This study aimed to explore the potential targets of the estrogen receptor (ER) signaling pathway and provide a meaningful way to treat DLBCL patients. Datasets were obtained from the Gene Expression Omnibus (GEO) to identify differentially expressed genes (DEGs). Representative gene sets estrogen receptor pathways, and growth regulatory pathways were identified based on Gene Set Enrichment Analysis (GSEA) analysis. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) were used for function and pathway analysis. STRING and Cytoscape were used to construct the interaction network, and the MCODE plug-in performed the module analysis. GEPIA, TCGA, and LOGpc databases were used for expression and predictive analysis. The Human Protein Atlas (HPA) database was used to analyze the protein expression levels, cBioPortal was used to explore genetic alterations, and ROC analysis and prognostic assessment were used to predict the diagnostic value of genes. Finally, BJAB cells were treated with ER inhibitor fulvestrant and specific shRNA, and the expression of hub genes was verified by RT-qPCR. We identified 81 overlapping DEGs and CDC6, CDC20, KIF20A, STIL, and TOP2A as novel biomarkers affecting the prognosis of DLBCL. In addition, the STAT and KRAS pathways are considered potential growth regulatory pathways. These results hold promise for new avenues for the treatment of DLBCL patients.

Head and Neck Squamous Cell Carcinoma Subtypes Based on Immunologic and Hallmark Gene Sets in Tumor and Non-tumor Tissues.

BackgroundNon-tumor tissue has a significant impact on the prognosis of head and neck squamous cell carcinoma (HNSCC). Previous studies for HNSCC have mainly focused on tumor tissue, greatly neglecting the role of non-tumor tissue. This study aimed to identify HNSCC subtypes and prognostic gene sets based on activity changes of immunologic and hallmark gene sets in tumor and adjacent non-tumor tissues to improve patient prognosis.MethodsIn the study, we used gene set variation analysis (GSVA) to estimate the relative enrichment of gene sets over the sample population, and identified relevant subtypes of HNSCC by Cox regression analysis and the non-negative matrix factorization (NMF) method. The representative gene sets were identified by calculating the differential enrichment score of gene sets between each of the two subgroups, intersecting them, and screening them using univariate Cox regression analysis. The least absolute shrinkage and selection operator (LASSO) regression analysis was used to screen out potential prognostic gene sets and establish a risk model. Finally, genes encompassed in each prognostic gene set were obtained and subjected to enrichment analysis and protein–protein interaction (PPI) in tumor and non-tumor tissues.ResultsWe identified three subtypes of HNSCC based on gene sets in tumor and non-tumor tissues, and patients with subtype 1 had a higher survival rate than subtypes 2 and 3. The subtypes were related to the survival status, pathological stage, and T stage of HNSCC patients. In total 450 differentially gene sets and 39 representative gene sets were obtained by calculating the differential enrichment score of gene sets between each of the two subgroups, intersecting them, and screening them using univariate Cox regression analysis. The prognostic model was constructed by LASSO regression analysis, including five prognostic gene sets. Kaplan-Meier analysis indicated that different risk groups and the five prognostic gene sets were associated with survival status in the model. Finally, enrichment analysis and PPI indicated that non-tumor and tumor tissues affect the prognosis of HNSCC patients in different ways.ConclusionIn conclusion, we provide a novel insight for rational treatment strategies and precise prognostic assessments based on tumor and adjacent non-tumor tissues, suggesting that more emphasis should be placed on changes in adjacent non-tumor and tumor tissues, rather than just the tumor itself.

Identification of new head and neck squamous cell carcinoma subtypes and development of a novel score system (PGSscore) based on variations in pathway activity between tumor and adjacent non-tumor samples

The influence of adjacent non-tumor tissue characteristics on patient outcomes in head and neck squamous cell carcinoma (HNSC) remains unclear. In HNSC, subtype identification is generally tumor-based, and most prognosis-related studies focus on a single gene set. Here, we performed Gene Set Variation Analysis to comprehensively evaluate variations in diverse gene sets in tumor and non-tumor samples and converted a gene-centric matrix into a pathway-centric model. Three different prognostic subtypes correlated with previously identified subgroups, clinicopathologic features, risk factors, and tumor microenvironment (TME) were identified using the non-negative matrix factorization method. We also screened 2 and 11 gene sets from nontumor and tumor tissues, respectively based on representative gene sets. Interestingly, genes from nontumor gene sets were associated with serotonin secretion and P2Y receptors, while genes from tumor gene sets were associated with immunity and inflammation. The PGSscore was constructed to predict outcomes and immunotherapy responses. Low- and high- PGSscore groups showed significant differences in clinicopathological characteristics and TME. Our analysis indicates that the non-tumor tissue is indispensable for prognosis. PGSscore provides new avenue to evaluate overall survival and immunotherapy responses, and our method based on pathway-centric models can be extended to other diseases.

Could microtubule inhibitors be the best choice of therapy in gastric cancer with high immune activity: mutant DYNC1H1 as a biomarker.

Immune checkpoint blockade (ICB) has achieved unprecedented breakthroughs in various cancers, including gastric cancer (GC) with high immune activity (MSI-H or TMB-H), yet clinical benefits from ICB were moderate. Here we aimed to identify the most appropriate drugs which can improve outcomes in GC. We firstly compared MSI-H and TMB-H GC samples with normal samples in TCGA-STAD cohort, respectively. After that, Connectivity Map database repurposed nine candidate drugs (CMap score < -90). Then, microtubule inhibitors (MTIs) were screened as the significant candidate drugs with their representative gene sets strongly enriched (p < 0.05) via GSEA. GDSC database validated higher activities of some MTIs in GC cells with MSI-H and TMB-H (p < 0.05). Furthermore, some MTIs activities were positively associated with mutant Dynein Cytoplasmic 1 Heavy Chain 1 (DYNC1H1) (p < 0.05) based on NCI-60 cancer cell line panel. DYNC1H1 was high frequently alteration in GC and was positively associated with TMB-H and MSI-H. Mutant DYNC1H1 may be accompanied with down-regulation of MTIs-related genes in GC or change the binding pocket to sensitize MTIs. Overall, this study suggested that some MTIs may be the best candidate drugs to treat GC with high immune activity, especially patients with DYNC1H1 mutated.

Identification of the association between HMMR expression and progression of hepatocellular carcinoma via construction of a co-expression network.

The aim of the present study was to identify key genes involved in the progression of hepatocellular carcinoma (HCC). According to the theory of the multistep process of hepatocarcinogenesis and weighted gene co-expression network analysis, hub genes associated with the progression of HCC were identified using the gene expression profiles of patients with normal to chronic hepatitis/cirrhosis and dysplastic nodules to HCC. An independent dataset was used to verify the association between hub gene and clinical phenotype. The diagnostic and prognostic value of hub genes regarding HCC were evaluated. Gene set enrichment analysis (GSEA) was performed to explore the function of hub genes. A co-expression gene module positively associated with HCC progression was identified. Combined with a protein-protein interaction (PPI) network, a total of 10 common hub genes common to both the module of interest and the PPI network were selected as hub genes. Hyaluronan mediated motility receptor (HMMR) was selected as the candidate gene and was significantly upregulated in HCC at the mRNA and protein expression levels. HMMR is a promising diagnostic biomarker for HCC, and is also associated with its progression. The expression of HMMR was positively correlated with HCC tumor grade, pathological stage, tumor stage and Ishak score. The expression of HMMR was an independent prognostic factor compared with clinicopathological features. Patients with high expression levels of HMMR exhibited a less favorable prognosis. GSEA identified 6 representative gene sets that were associated with cancer. Overall, HMMR may serve an important role in HCC and may have potential as a biomarker of HCC diagnosis and progression.

Graph Algorithms for Condensing and Consolidating Gene Set Analysis Results

Gene set analysis plays a critical role in the functional interpretation of omics data. Although this is typically done for one omics experiment at a time, there is an increasing need to combine gene set analysis results from multiple experiments performed on the same or different omics platforms, such as in multi-omics studies. Integrating results from multiple experiments is challenging, and annotation redundancy between gene sets further obscures clear conclusions. We propose to use a weighted set cover algorithm to reduce redundancy of gene sets identified in a single experiment. Next, we use affinity propagation to consolidate similar gene sets identified from multiple experiments into clusters and to automatically determine the most representative gene set for each cluster. Using three examples from over representation analysis and gene set enrichment analysis, we showed that weighted set cover outperformed a previously published set cover method and reduced the number of gene sets by 52-77%. Focusing on overlapping genes between the list of input genes and the enriched gene sets in over-representation analysis and leading-edge genes in gene set enrichment analysis further reduced the number of gene sets. A use case combining enrichment analysis results from RNA-Seq and proteomics data comparing basal and luminal A breast cancer samples highlighted the known difference in proliferation and DNA damage response. Finally, we used these algorithms for a pan-cancer survival analysis. Our analysis clearly revealed prognosis-related pathways common to multiple cancer types or specific to individual cancer types, as well as pathways associated with prognosis in different directions in different cancer types. We implemented these two algorithms in an R package, Sumer, which generates tables and static and interactive plots for exploration and publication. Sumer is publicly available at https://github.com/bzhanglab/sumer.

Feature clustering based support vector machine recursive feature elimination for gene selection

In a DNA microarray dataset, gene expression data often has a huge number of features(which are referred to as genes) versus a small size of samples. With the development of DNA microarray technology, the number of dimensions increases even faster than before, which could lead to the problem of the curse of dimensionality. To get good classification performance, it is necessary to preprocess the gene expression data. Support vector machine recursive feature elimination (SVM-RFE) is a classical method for gene selection. However, SVM-RFE suffers from high computational complexity. To remedy it, this paper enhances SVM-RFE for gene selection by incorporating feature clustering, called feature clustering SVM-RFE (FCSVM-RFE). The proposed method first performs gene selection roughly and then ranks the selected genes. First, a clustering algorithm is used to cluster genes into gene groups, in each which genes have similar expression profile. Then, a representative gene is found to represent a gene group. By doing so, we can obtain a representative gene set. Then, SVM-RFE is applied to rank these representative genes. FCSVM-RFE can reduce the computational complexity and the redundancy among genes. Experiments on seven public gene expression datasets show that FCSVM-RFE can achieve a better classification performance and lower computational complexity when compared with the state-the-art-of methods, such as SVM-RFE.

A predictive signature gene set for discriminating active from latent tuberculosis in Warao Amerindian children.

BackgroundTuberculosis (TB) continues to cause a high toll of disease and death among children worldwide. The diagnosis of childhood TB is challenged by the paucibacillary nature of the disease and the difficulties in obtaining specimens. Whereas scientific and clinical research efforts to develop novel diagnostic tools have focused on TB in adults, childhood TB has been relatively neglected. Blood transcriptional profiling has improved our understanding of disease pathogenesis of adult TB and may offer future leads for diagnosis and treatment. No studies applying gene expression profiling of children with TB have been published so far.ResultsWe identified a 116-gene signature set that showed an average prediction error of 11% for TB vs. latent TB infection (LTBI) and for TB vs. LTBI vs. healthy controls (HC) in our dataset. A minimal gene set of only 9 genes showed the same prediction error of 11% for TB vs. LTBI in our dataset. Furthermore, this minimal set showed a significant discriminatory value for TB vs. LTBI for all previously published adult studies using whole blood gene expression, with average prediction errors between 17% and 23%. In order to identify a robust representative gene set that would perform well in populations of different genetic backgrounds, we selected ten genes that were highly discriminative between TB, LTBI and HC in all literature datasets as well as in our dataset. Functional annotation of these genes highlights a possible role for genes involved in calcium signaling and calcium metabolism as biomarkers for active TB. These ten genes were validated by quantitative real-time polymerase chain reaction in an additional cohort of 54 Warao Amerindian children with LTBI, HC and non-TB pneumonia. Decision tree analysis indicated that five of the ten genes were sufficient to classify 78% of the TB cases correctly with no LTBI subjects wrongly classified as TB (100% specificity).ConclusionsOur data justify the further exploration of our signature set as biomarkers for potential childhood TB diagnosis. We show that, as the identification of different biomarkers in ethnically distinct cohorts is apparent, it is important to cross-validate newly identified markers in all available cohorts.

De novo analysis of transcriptome dynamics in the migratory locust during the development of phase traits.

Locusts exhibit remarkable density-dependent phenotype (phase) changes from the solitary to the gregarious, making them one of the most destructive agricultural pests. This phenotype polyphenism arises from a single genome and diverse transcriptomes in different conditions. Here we report a de novo transcriptome for the migratory locust and a comprehensive, representative core gene set. We carried out assembly of 21.5 Gb Illumina reads, generated 72,977 transcripts with N50 2,275 bp and identified 11,490 locust protein-coding genes. Comparative genomics analysis with eight other sequenced insects was carried out to indentify the genomic divergence between hemimetabolous and holometabolous insects for the first time and 18 genes relevant to development was found. We further utilized the quantitative feature of RNA-seq to measure and compare gene expression among libraries. We first discovered how divergence in gene expression between two phases progresses as locusts develop and identified 242 transcripts as candidates for phase marker genes. Together with the detailed analysis of deep sequencing data of the 4th instar, we discovered a phase-dependent divergence of biological investment in the molecular level. Solitary locusts have higher activity in biosynthetic pathways while gregarious locusts show higher activity in environmental interaction, in which genes and pathways associated with regulation of neurotransmitter activities, such as neurotransmitter receptors, synthetase, transporters, and GPCR signaling pathways, are strongly involved. Our study, as the largest de novo transcriptome to date, with optimization of sequencing and assembly strategy, can further facilitate the application of de novo transcriptome. The locust transcriptome enriches genetic resources for hemimetabolous insects and our understanding of the origin of insect metamorphosis. Most importantly, we identified genes and pathways that might be involved in locust development and phase change, and may thus benefit pest management.

Genome SEGE: A database for 'intronless' genes in eukaryotic genomes

BackgroundA number of completely sequenced eukaryotic genome data are available in the public domain. Eukaryotic genes are either 'intron containing' or 'intronless'. Eukaryotic 'intronless' genes are interesting datasets for comparative genomics and evolutionary studies. The SEGE database containing a collection of eukaryotic single exon genes is available. However, SEGE is derived using GenBank. The redundant, incomplete and heterogeneous qualities of GenBank data are a bottleneck for biological investigation in comparative genomics and evolutionary studies. Such studies often require representative gene sets from each genome and this is possible only by deriving specific datasets from completely sequenced genome data. Thus Genome SEGE, a database for 'intronless' genes in completely sequenced eukaryotic genomes, has been constructed.Availability: DescriptionEukaryotic 'intronless' genes are extracted from nine completely sequenced genomes (four of which are unicellular and five of which are multi-cellular). The complete dataset is available for download. Data subsets are also available for 'intronless' pseudo-genes. The database provides information on the distribution of 'intronless' genes in different genomes together with their length distributions in each genome. Additionally, the search tool provides pre-computed PROSITE motifs for each sequence in the database with appropriate hyperlinks to InterPro. A search facility is also available through the web server.ConclusionsThe unique features that distinguish Genome SEGE from SEGE is the service providing representative 'intronless' datasets for completely sequenced genomes. 'Intronless' gene sets available in this database will be of use for subsequent bio-computational analysis in comparative genomics and evolutionary studies. Such analysis may help to revisit the original genome data for re-examination and re-annotation.