A transcriptome-based classifier to determine molecular subtypes in medulloblastoma

Polycystic ovary syndrome (PCOS) is one of the most common metabolic and reproductive endocrinopathies. However, few studies have tried to develop a diagnostic model based on gene biomarkers. In this study, we applied a computational method by combining two machine learning algorithms, including random forest (RF) and artificial neural network (ANN), to identify gene biomarkers and construct diagnostic model. We collected gene expression data from Gene Expression Omnibus (GEO) database containing 76 PCOS samples and 57 normal samples; five datasets were utilized, including one dataset for screening differentially expressed genes (DEGs), two training datasets, and two validation datasets. Firstly, based on RF, 12 key genes in 264 DEGs were identified to be vital for classification of PCOS and normal samples. Moreover, the weights of these key genes were calculated using ANN with microarray and RNA-seq training dataset, respectively. Furthermore, the diagnostic models for two types of datasets were developed and named neuralPCOS. Finally, two validation datasets were used to test and compare the performance of neuralPCOS with other two set of marker genes by area under curve (AUC). Our model achieved an AUC of 0.7273 in microarray dataset, and 0.6488 in RNA-seq dataset. To conclude, we uncovered gene biomarkers and developed a novel diagnostic model of PCOS, which would be helpful for diagnosis.

1 Abstract Understanding the expression changes that come with age is an important step in understanding the ageing process as a whole. By combining such transcriptomic data with other sources of information, for instance protein-protein interaction (PPI) data, it is possible to make inferences about the functional changes that occur with age. To address this, we conducted a meta-analysis on 127 publicly available microarray and RNA-Seq datasets from mice, rats and humans, to identify genes that are commonly differentially expressed with age in mammals. We also conducted analyses on subsets of these datasets, to produce transcriptomic signatures for brain, heart and muscle tissues, all of which are important tissues in the pathophysiology of ageing. This approach identified the transcriptomic signatures of the ageing system, as well as brain, heart and muscle tissues. We then applied enrichment analysis and machine learning to functionally describe those signatures. This revealed a typical ageing signature including the overexpression of immune and stress response genes and the underexpression of metabolic and developmental genes. Further analysis of the ageing expression signatures revealed that genes differentially expressed with age tend to be broadly expressed across tissues, rather than be tissue-specific, and that the ageing expression signatures (particularly the overexpressed signatures) of the whole system, brain and muscle tend to include genes that are central in PPI networks. We also show that genes underexpressed in the brain are highly central in a co-expression map, suggesting that underexpression of these genes may play a part in cognitive ageing. In sum, we show numerous functional similarities between the ageing transcriptomes of these important tissues, a broad non-specific expression pattern in genes differentially expressed with age, along with altered network properties of these genes in both a PPI and co-expression network.

Triple-negative breast cancer (TNBC) accounts for 15-20% of primary BC. This heterogeneous disease is the most aggressive BC subtype. TNBC is characterized by young age at onset, high-grade tumors, high risk of early recurrence and poor prognosis. Until recently, TNBC patients were considered as one group and all patients were treated with conventional chemotherapy. With the emergence of new therapies targeting distinct biological pathways, personalized medicine is now accessible by defining finer TNBC subtypes, e.g. molecular subtypes, and identifying their specific targetable pathways. Several studies based on transcriptomics have identified three to six TNBC molecular subtypes. Today, there is a clear consensus on three TNBC robust subtypes: luminal androgen receptor (LAR), basal-like immune-activated (BLIA) and basal-like immune-suppressed (BLIS). The debate about the robustness of other subtypes is still open, notably mesenchymal-like (M) and mesenchymal stem-like (MSL or MES). The goal of this study was to establish a robust molecular classification of TNBC that allow to determine the best treatment options. To optimize this process, we harvested as much transcriptomic data as possible. Available clinicopathological data linked to these data were also imported. To our knowledge, this is the largest study aimed at subtyping TNBC; it exploits a total of 1942 TNBC patient data collected from 19 datasets: 1243 samples from microarray technology (14 datasets) and 699 from RNAseq technology (five datasets including a new one from our institution [GSE225002]). We performed independent unsupervised analyses on these two large cohorts with consensus k-means clustering including selection of optimal number of clusters and optimal number of variables. Both consensus clustering analyses separated TNBC microarray and RNAseq TNBC cohorts into four optimal clusters. These clusters were then characterized using clinicopathological features, gene ontology enrichment analyses (GOEA), gene set enrichment analysis (GSEA), 70 gene expression signatures (GES) and expression of 47 immune checkpoint genes. We concluded that TNBC are composed of four robust subtypes: LAR, mesenchymal-like immune-activated (MLIA), BLIA and BLIS. Importantly, we propose that MLIA acronym should replace M or MSL, because it highlights the two main biological characteristics of this subtype, i.e. mesenchymal-like features and a high IR comparable to that of BLIA. Furthermore, we showed that each subtype shares common biological characteristics with another one: LAR and MLIA (androgen receptor pathway activation); MLIA and BLIA (immune activation, notably high levels of immune checkpoints); BLIA and BLIS (basal-like features); BLIS and LAR (immune suppression). As regards microarray cohort, LAR included 19.23% (n = 239) of TNBC, MLIA 17.38% (n = 216), BLIA 32.50% (n = 404) and BLIS 30.89% (n = 384). Concerning RNAseq cohort, LAR included 19.60% (n = 137) of TNBC, MLIA 17.31% (n = 121), BLIA 28.47% (n = 199) and BLIS 34.62% (n = 242). Distributions of patients among the four clusters in the two cohorts were similar (p = 0.3583). Distributions were also similar between studies for microarray cohort (p = 0.2479) and RNAseq cohort (p = 0.1775). Pooled data (n = 1942 patients) gave the following distribution: 19.36% for LAR; 17.35% for MLIA; 31.06% for BLIA; 32.23% for BLIS, confirming that TNBC are largely but not exclusively basal-like (63.29%). In brief, according to the TNBC subtype molecular characterisation, anti-androgen therapy could be prescribed for LAR, immune checkpoint inhibitors for MLIA and BLIA, alone or in combination, and TAM-focused and antineurogenic therapies for BLIS. As an extension of this work, TNBC transcriptomic data associated with their clinicopathological features and TNBC subtype annotations were included in bc-GenExMiner v5.0 web tool (http://bcgenex.ico.unicancer.fr). Citation Format: Mario Campone, Pascal Jézéquel, Fadoua Ben Azzouz, Wilfried Gouraud, Agnes Bassevile, Bertrand Michel, Hamza Lasla, Jean Sebastien FRENEL. Mesenchymal-like immune-activated: the 4th robust triple-negative breast cancer molecular subtype [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO2-15-05.

Prostate cancer (PCa) is the second most common cancer for men worldwide.Nevertheless, prevention, treatment, and patient health care improved the probability for men to survive PCa.However, current screenings and imprecise prognostic tests lead to a high number of overdiagnosed and overtreated patients accompanied by adverse effects and health care burden.Therefore, more specific prognostic and predictive tests are necessary to distinguish between benign and aggressive tumor to be able to adapt therapy accordingly.As a heterogeneous disease, PCa forms several subtypes.Its most prevalent form is characterized by the TMPRSS2-ERG (T2E) gene fusion, which appears in around 50% of PCa.Differentiating etiopathology, progression driving pathways, and disease outcome of PCa subtypes make their treatment challenging.However, many established PCa test do not distinguish between subtypes, so far, possibly affecting outcome prediction and treatment decisions.Recent advances in high throughput technologies allow the extensive generation of genomic and transcriptomic data and pioneered personalized medicine empowering clinical diagnostics and risk prediction.Consequently, identifying suitable biomarkers via genome-wide association studies (GWAS) and gene expression analysis is crucial for the development of gene and variant based prognostic as well as predictive tests.This thesis assessed the contribution of somatic mutations and germline variation in PCa progression.It emphasized the molecular differences between PCa subtypes by investigating their gene-signatures associated with metastasis, affirming the importance of considering the T2E-status of a patient in research studies as well as in clinical settings.Additionally, subtype specific biomarkers were identified showing that their prognostic value decisively depends on the T2E-status.Furthermore, a risk locus (7q31.33)was detected that harbored five potential tag single nucleotide polymor-xii Summary phisms (SNPs) associated with aggressive, high-grade PCa and partially altering GRM8 expression, thus, implying an influence of germline variants on PCa progression.These findings were obtained from a meta-analysis of multiple GWAS and a gene expression analysis of PCa subtypes:GWAS testing for germline association with Gleason grade (GG) were performed on three PCa cohorts including 1,098 Central European men.While two of the cohorts were whole exon/genome sequenced (TCGA-PRAD, ICGC-CA), the third, own cohort (PCGA LMU ) was genotyped on a microarray.The results were combined in a meta-analysis.One genome-wide significant SNP (rs12537032) was found at the risk locus chr7q31.33,which is associated with aggressive PCa (high GG) and comprises the GRM8 gene.With fine-mapping, five tag SNPs (rs12537032, rs74999840, rs1910298267, rs76326523, and rs73451279) from two independent signals on the same haploblock could be detected, which were in linkage disequilibrium with each other, respectively.The risk/minor allele of these tag SNPs increased the risk for high-grade PCa up to a factor of three.Located in intronic regions of GRM8 only the risk allele G of rs73451279 was associated with lower gene expression (via eQTL).Variant effect prediction indicated that the risk allele may induce nonsense-mediated mRNA decay, affecting PCa progression.In survival analysis, however, neither the genotype nor GRM8 expression was associated with worse PCa outcome.Conditional analysis showed that rs73451279 might not be the only functional variant at this locus affecting mechanisms that drive the development of high-grade PCa.Based on comprehensive transcriptomic and matched clinicopathological data of two discovery cohorts (n = 783), the gene expression profiles of T2E-positive andnegative PCa were independently investigated to compare metastasis associated genesignatures regarding their T2E-status.With gene set enrichment analyses distinct genesignatures characterizing T2E-positive and -negative tumors were detected.Genes frequently involved in these functional gene-signatures were further investigated in a validation cohort (n = 272).Beside being associated with metastasis, five genes (ASPN, BGN, COL1A1, RRM2, and TYMS) were also identified to be associated with event-free survival.In exclusively T2E-negative tumors, their overexpression was significantly associated with worse outcome.RRM2 and TYMS were both evaluated in another validation cohort (n = 135) by immunohistochemistry.Moreover, some of these subtype spe-xiii cific biomarkers did enhance the Gleason score as current clinicopathological predictor in T2E-negative PCa.In contrast, these observations did not apply for T2E-positive tumors for which no biomarker could be found in this study.Taken together, these findings strongly suggest considering molecular subtypes in combination with the application of prognostic biomarkers to improve outcome prediction in PCa.

Background: Recent studies have identified several molecular subtypes of lung adenocarcinoma (LUAD) that have different prognoses to help predict the efficacy of immunotherapy. However, the prognostic prediction is less than satisfactory. Alterations in intracellular copper levels may affect the tumor immune microenvironment and are linked to cancer progression. Previous studies have identified some genes related to cuproptosis. The characteristics of the cuproptosis molecular subtypes have not been thoroughly studied in LUAD. Methods: The transcriptomic data and clinical information of 632 LUAD patients were used to investigate the LUAD molecular subtypes that are associated with the cuproptosis-related genes (CRGs), the tumor immune microenvironment, and stemness. The cuproptosis score was constructed using univariate Cox regression and the minor absolute shrinkage and selection operator (LASSO) to quantify the prognostic characteristics. Results: Three different molecular subtypes related to cuproptosis, with different prognoses, were identified in LUAD. Cluster A had the highest cuproptosis score and the worst prognosis. Patients in the high cuproptosis score group had a higher somatic mutation frequency and stemness scores. Patients in the low cuproptosis score group had more immune infiltration and better prognosis. Conclusion: Molecular subtypes of LUAD based on CRGs reflect the differences in LUAD patients. The cuproptosis score can be used as a promising biomarker, which is of great significance to distinguish the relationship between cuproptosis and the immune microenvironment. The cuproptosis signature based on the cuproptosis score and clinical characteristics of individual patients will be useful for guiding immunotherapy in LUAD.

Background: Isocitrate dehydrogenase (IDH) mutant is one of the most robust and important genetic aberrations in glioma. However, the underlying regulation mechanism of long non-coding RNA (lncRNA) in IDH mutant glioma has not been systematically portrayed.Methods:In this work, 775 IDH mutant glioma samples with transcriptome data, including 167 samples from the Chinese Glioma Genome Atlas (CGGA) RNAseq dataset, 390 samples from The Cancer Genome Atlas (TCGA) dataset, 79 samples from GSE16011 dataset, and 139 samples from CGGA microarray dataset, were enrolled. R language and GraphPad Prism software were applied for the statistical analysis and graphical work.Results: By comparing the differentially lncRNA genes between IDH mutant and IDH wild-type glioma samples, a four-lncRNA (JAG1, PVT1, H19, and HAR1A) signature was identified in IDH mutant glioma patients. The signature model was established based on the expression level and the regression coefficient of the four lncRNA genes. IDH mutant glioma samples could be successfully stratified into low-risk and high-risk groups in CGGA RNAseq, TCGA, GSE16011, and CGGA microarray databases. Meanwhile, multivariate Cox analysis showed that the four-lncRNA signature was an independent prognostic biomarker after adjusting for other clinicopathologic factors. Moreover, the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses revealed that the immune response and cellular metabolism were significantly associated with the four-lncRNA risk signature.Conclusion: Taken together, the four-lncRNA risk signature was identified as a novel prognostic marker for IDH mutant glioma patients and may potentially lead to improvements in the lives of glioma patients.

High-throughput RNA sequencing (RNA-seq) has become an instrumental assay for the analysis of multiple aspects of an organism's transcriptome. Further, the analysis of a biological specimen's associated microbiome can also be performed using RNA-seq data and this application is gaining interest in the scientific community. There are many existing bioinformatics tools designed for analysis and visualization of transcriptome data. Despite the availability of an array of next generation sequencing (NGS) analysis tools, the analysis of RNA-seq data sets poses a challenge for many biomedical researchers who are not familiar with command-line tools. Here we present RNA CoMPASS, a comprehensive RNA-seq analysis pipeline for the simultaneous analysis of transcriptomes and metatranscriptomes from diverse biological specimens. RNA CoMPASS leverages existing tools and parallel computing technology to facilitate the analysis of even very large datasets. RNA CoMPASS has a web-based graphical user interface with intrinsic queuing to control a distributed computational pipeline. RNA CoMPASS was evaluated by analyzing RNA-seq data sets from 45 B-cell samples. Twenty-two of these samples were derived from lymphoblastoid cell lines (LCLs) generated by the infection of naïve B-cells with the Epstein Barr virus (EBV), while another 23 samples were derived from Burkitt's lymphomas (BL), some of which arose in part through infection with EBV. Appropriately, RNA CoMPASS identified EBV in all LCLs and in a fraction of the BLs. Cluster analysis of the human transcriptome component of the RNA CoMPASS output clearly separated the BLs (which have a germinal center-like phenotype) from the LCLs (which have a blast-like phenotype) with evidence of activated MYC signaling and lower interferon and NF-kB signaling in the BLs. Together, this analysis illustrates the utility of RNA CoMPASS in the simultaneous analysis of transcriptome and metatranscriptome data. RNA CoMPASS is freely available at http://rnacompass.sourceforge.net/.

BackgroundHeart failure (HF) is the end stage of various cardiovascular diseases with a high mortality rate. Novel diagnostic and therapeutic biomarkers for HF are urgently required. Our research aims to identify HF-related hub genes and regulatory networks using bioinformatics and validation assays.MethodsUsing four RNA-seq datasets in the Gene Expression Omnibus (GEO) database, we screened differentially expressed genes (DEGs) of HF using Removal of Unwanted Variation from RNA-seq data (RUVSeq) and the robust rank aggregation (RRA) method. Then, hub genes were recognized using the STRING database and Cytoscape software with cytoHubba plug-in. Furthermore, reliable hub genes were validated by the GEO microarray datasets and quantitative reverse transcription polymerase chain reaction (qRT-PCR) using heart tissues from patients with HF and non-failing donors (NFDs). In addition, R packages “clusterProfiler” and “GSVA” were utilized for enrichment analysis. Moreover, the transcription factor (TF)–DEG regulatory network was constructed by Cytoscape and verified in a microarray dataset.ResultsA total of 201 robust DEGs were identified in patients with HF and NFDs. STRING and Cytoscape analysis recognized six hub genes, among which ASPN, COL1A1, and FMOD were confirmed as reliable hub genes through microarray datasets and qRT-PCR validation. Functional analysis showed that the DEGs and hub genes were enriched in T-cell-mediated immune response and myocardial glucose metabolism, which were closely associated with myocardial fibrosis. In addition, the TF–DEG regulatory network was constructed, and 13 significant TF–DEG pairs were finally identified.ConclusionOur study integrated different RNA-seq datasets using RUVSeq and the RRA method and identified ASPN, COL1A1, and FMOD as potential diagnostic biomarkers for HF. The results provide new insights into the underlying mechanisms and effective treatments of HF.

CO0005 C-C CHEMOKINE RECEPTOR TYPE 5 AND ITS LIGANDS CCL4, 8 AND 11 CAN LINK COVID-19, RHEUMATOID ARTHRITIS AND HYDROXYCHLOROQUINE

Lipid-soluble signaling molecule-related genes (LSMRGs) are critical in various tumors, but their role in gastric cancer (GC) prognosis and therapy remains unclear. Using transcriptomic data, this study analyzed LSMRG expression patterns, molecular subtypes, prognostic significance, and immune-microenvironment interactions in GC to identify new prognostic biomarkers and support precision medicine approaches. Using GC data from TCGA and GEO, LSMRGs from Genecard were analyzed. Unsupervised clustering defined LSMRG-based subtypes. Differentially expressed LSMRGs were identified by intersecting tumor-normal DEGs. A prognostic risk-score model was built via univariate Cox, LASSO, and multivariate Cox analyses, with model genes validated by qRT-PCR in cell lines. Comprehensive transcriptomic analyses included nomogram development, gene enrichment, immune infiltration, somatic mutations, and drug sensitivity. LSMRG-based clustering identified two patient subtypes with distinct survival and immunotherapy responses. From 83 differentially expressed LSMRGs, a 6-gene prognostic risk-score model was constructed, validated as an independent prognostic factor. Model gene expression was confirmed via qRT-PCR. The risk score accurately predicted 1-, 3-, and 5-year survival. High- and low-risk groups exhibited differential enrichment in pathways including neuroactive ligand-receptor interaction and hormone signaling. The high-risk group had a higher mutation burden, lower immune infiltration, and distinct drug sensitivity profiles compared to the low-risk group. This study delineates the expression landscape of LSMRGs in GC and clarifies their associations with molecular subtypes, prognosis, and immune regulation. The findings provide novel prognostic biomarkers and a molecular basis for targeting lipid-soluble signaling pathways in GC therapy.

Molecular subtyping and functional validation of TTK, TPX2, UBE2C, and LRP8 in sensitivity of TNBC to paclitaxel

Impact of renal cell carcinoma molecular subtypes on immunotherapy and targeted therapy outcomes

High-throughput sequencing has provided the capacity of broad virus detection for both known and unknown viruses in a variety of hosts and habitats. It has been successfully applied for novel virus discovery in many agricultural crops, leading to the current drive to apply this technology routinely for plant health diagnostics. For this, efficient and precise methods for sequencing-based virus detection and discovery are essential. However, both existing alignment-based methods relying on reference databases and even more recent machine learning approaches are not efficient enough in detecting unknown viruses in RNAseq datasets of plant viromes. We present VirHunter, a deep learning convolutional neural network approach, to detect novel and known viruses in assemblies of sequencing datasets. While our method is generally applicable to a variety of viruses, here, we trained and evaluated it specifically for RNA viruses by reinforcing the coding sequences’ content in the training dataset. Trained on the NCBI plant viruses data for three different host species (peach, grapevine, and sugar beet), VirHunter outperformed the state-of-the-art method, DeepVirFinder, for the detection of novel viruses, both in the synthetic leave-out setting and on the 12 newly acquired RNAseq datasets. Compared with the traditional tBLASTx approach, VirHunter has consistently exhibited better results in the majority of leave-out experiments. In conclusion, we have shown that VirHunter can be used to streamline the analyses of plant HTS-acquired viromes and is particularly well suited for the detection of novel viral contigs, in RNAseq datasets.

As the most common malignant pediatric brain cancer, medulloblastoma (MB) accounts for around 20% of all pediatric central nervous system (CNS) neoplasms. MB includes a complex array of distinct molecular subtypes characterized by morphological, molecular, and genetic alterations, mainly including SHH, WNT, Group3 and Group4. Although the overall cure rates for MB are around 70%, the survival rate varies significantly across different molecular subtypes. Conventional experimental methods for identifying MB subtypes are usually costly, time-consuming, and labor-intensive. Accurate identification of MB subtypes enables improved downstream risk stratification and tailored therapeutic treatment design. Previous studies indicate that identifying MB subtypes based on transcriptome data is an efficient and feasible way. However, their performance may be poor due to limited cohorts and severe batch effects if combining data in different sequencing platforms and various data sources. To overcome these limitations, we propose a novel computational approach based on intra-sample subtype-specific ranked genes (iS3RGs) analysis to effectively integrate heterogeneous transcriptomic data with batch effects for accurate identification of MB subtypes across diverse cohorts. For a baseline microarray dataset comprising 763 samples as training part, we specifically converted the absolute gene expression to the intra-sample rank values to capture reliable gene patterns by reducing bias from fluctuations and measurement errors. By comparing gene pairs within cohorts rather than across cohorts, we could avoid batch effects and successfully identified those gene pairs that significantly contributed to the identification of MB subtypes. Subsequently, we performed differential ranking analysis for iS3RGs identification. These genes were organized into gene ranking ratios and the subtype-specific gene ratios were filtered using Fisher’s exact test and LASSO regression, based on which 1035 gene ratios (corresponding to 738 unique genes) were identified. Then by testing on 14 independent datasets, our proposed method significantly outperformed state-of-the-art MB subtyping approaches with a median accuracy of 99.02%. We believe that our proposed approach has direct positive impacts on downstream MB risk stratification and tailored treatment design. We also expect that our proposed approach has great potential to address many biomedical problems with batch effects for various types of pediatric, adolescent, and young adult (AYA) cancers. Citation Format: Mengtao Sun, Jieqiong Wang, Shibiao Wan. iS3RGs: Discriminating diverse medulloblastoma subtypes by leveraging heterogenous transcriptome data with batch effects [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 5022.

Taking a byte out of big data