Integrative Machine Learning Approaches to Identify and Validate Gene Biomarkers for Early Detection of Hepatocellular Carcinoma DOI Creative Commons

Mahati Munikoti Srikantamur,

Parneet Kaur,

Eckart Bindewald

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 30, 2024

Abstract Hepatocellular carcinoma (HCC) is among the leading causes of cancer-related deaths worldwide, and prognosis poor if disease detected at advanced stages. There an urgent need for early diagnostic biomarkers to facilitate timely interventions. Current methods, such as liver function tests (LFTs), alpha-fetoprotein (AFP) panels, imaging techniques like magnetic resonance (MRI) ultrasound, lack specificity HCC do not provide a comprehensive prognosis. This study proposes machine learning (ML) based approach identifying using RNA-sequencing (RNA-seq) data. We analyzed publicly available RNA-seq datasets from Gene Expression Omnibus (GEO), UCSC Xena, GEO Experiments Interactive Navigator (GREIN). In this study, we performed various feature selection methods ML with Random Forest (RF) model, achieving best performance in predicting top most significantly important genes. Bioinformatics tools, including Search Tool Retrieval Interacting Genes/Proteins (STRING), Ontology (GO), DAVID (Database Annotation, Visualization, Integrated Discovery), Human Protein Atlas (HPA), Comparative Toxicogenomics Database (CTD) were used validation. Through our analysis, identified six potential early-detection gene HCC: CDKN3, LIFR, MKI67, TOP2A, SLC5A1, VIPR1.

Language: Английский

A comprehensive pan-cancer analysis of RNF187 in human tumors DOI Creative Commons
Xuezhong Zhang, Xuebin Zhang, Tonggang Liu

et al.

Discover Oncology, Journal Year: 2025, Volume and Issue: 16(1)

Published: Jan. 13, 2025

RING Finger 187 (RNF187) has recently emerged as a potential contributor to tumorigenesis. However, comprehensive pan-cancer analysis of RNF187 in human tumors not been undertaken until now. Our study aims investigate expression across 33 different types tumors, utilizing data from the TCGA and GTEx databases. The revealed significant upregulation 27 cancers, contrasting with only low LAML, no statistical differences OV SARC. Notably, discernible associations were identified between prognosis cancer patients. investigation also unveiled correlations RNA modification various types. Further exploration indicated positive correlation levels presence cancer-associated fibroblasts (CAFs) numerous tumor Additionally, exhibited majority immune inhibitory stimulatory genes, well chemokines, receptors, MHC molecules, immunoinhibitors, immunostimulators cancers. findings highlighted Tumor Mutational Burden (TMB), Microsatellite Instability (MSI) Homologous Recombination Deficiency (HRD) specific tumors. Finally, showed association five genes (ALKBH4, FAM134A, MLST8, SANP47 TMEM9) GO enrichment KEGG pathway analyses suggested that may play role pathogenesis through processes such "bounding membrane organelle," "macroautophagy," "proton-transporting V-type ATPase complex," "autophagy," "Ubiquitin mediated proteolysis," "Ferroptosis," "Phagosome," etc. inaugural provide profound understanding tumorigenesis diverse

Language: Английский

Citations

1
Comprehensive pan-cancer analysis of ENOPH1 in human tumors DOI Creative Commons
Xuezhong Zhang, Ning Li,

Ting‐Ting Chu

et al.

Discover Oncology, Journal Year: 2025, Volume and Issue: 16(1)

Published: Feb. 15, 2025

ENOPH1 (Enolase-phosphatase 1), a member of the HAD-like hydrolase superfamily, has been linked to range physiological conditions, including neurological disorders. However, its involvement in tumorigenesis remains underexplored. This study is first conduct pan-cancer analysis ENOPH1, aiming elucidate role multiple cancers through various bioinformatics platforms. We conducted thorough using data from UCSC databases. expression tumor and normal tissues was evaluated R language software. Survival analyses, genetic alterations, RNA modifications were assessed GEPIA2 cBioPortal The relationships between immune infiltration, mutational burden (TMB), microsatellite instability (MSI), homologous recombination deficiency (HRD) examined TIMER2 ENOPH1-related gene enrichment performed STRING databases, followed by Gene Ontology (GO) KEGG pathway analyses. significantly upregulated cancers, ACC, BLCA, BRCA, COAD. High associated with poor overall survival (OS) such as KICH, LIHC, BRCA LUAD. disease specific (DSS) MESO. Genetic alterations primarily mutations deep deletions, identified UCEC, OV. showed significant correlations (m1A, m5C, m6A), checkpoints, modulators across cancer types. positively correlated TMB, MSI, HRD like STAD. Furthermore, revealed that interacts proteins involved critical pathways AMPK, Hippo, PI3K-AKT, suggesting progression. reveals ENOPH1's potential prognostic biomarker key signaling cancers. Our findings provide new insights into highlight therapeutic target treatment.

Language: Английский

Citations

0
Comprehensive analysis of TMEM9 in human tumors DOI Creative Commons
Xuezhong Zhang, Lele Zhou,

Shumin Wei

et al.

Discover Oncology, Journal Year: 2025, Volume and Issue: 16(1)

Published: March 7, 2025

TMEM9, a transmembrane protein, has emerged as significant player in tumor progression, yet its comprehensive role across various cancers remains unclear. This study investigates the expression patterns, genetic alterations, immune associations, and prognostic implications of TMEM9 multiple cancer types using large-scale bioinformatics approaches. was analyzed normal tissues data from TCGA GTEx databases, with protein verified via CPTAC datasets. The impact assessed overall survival (OS) disease-free (DFS) analyses cancers. Genetic including mutation copy number were explored cBioPortal platform. We also examined DNA methylation, RNA modifications, infiltration correlations tools, TIMER2 UALCAN. TMEM9's relationship mutational burden (TMB) microsatellite instability (MSI) analyzed, gene enrichment performed STRING database GO/KEGG pathway analyses. significantly overexpressed several cancers, ACC, BLCA, BRCA, CHOL, COAD, GBM, others. Elevated correlated worse OS DFS CESC, KICH, UVM, additional types. predominantly amplifications, frequent LIHC, UCEC. methylation analysis revealed hypermethylation tumors like HNSC KIRC, while modification showed associations m6A m1A-related genes. strongly infiltration, particularly cancer-associated fibroblasts THYM HNSC. Positive between TMB/MSI observed suggesting genomic instability. Enrichment identified involvement pathways related to endoplasmic reticulum, Wnt signaling, processing. plays crucial influencing expression, modulation, These findings highlight potential biomarker therapeutic target

Language: Английский

Citations

0
Pan-cancer bioinformatics analysis of TIPRL in human tumors DOI Creative Commons
Xuezhong Zhang, Hao Xue, Yuanyuan Lv

et al.

Discover Oncology, Journal Year: 2025, Volume and Issue: 16(1)

Published: March 15, 2025

The TOR signaling pathway regulator-like (TIPRL) gene plays a multifaceted role in cancer, yet its pan-cancer profile remains underexplored. This study investigates TIPRL expression across multiple cancers and associations with survival, genetic alterations, immune infiltration, functional pathways, providing insights into TIPRL's as potential prognostic therapeutic target. significance tumor types were analyzed using TCGA_GTEx CPTAC data R software platforms like GEPIA2 UALCAN. Genetic alterations 3D structures evaluated through cBioPortal. Associations RNA modifications, checkpoints, cell TMB, MSI, HRD, enriched pathways assessed via STRING databases, employing survival analysis, ssGSEA, enrichment analyses. was elevated most cancers, significant stage-specific observed KICH, KIRP, LUSC. High correlated worse overall ACC, BRCA, HNSC, LIHC, MESO, suggesting prognosis. analysis identified amplifications the main alteration, varied clinical relevance cancers. modifications TIPRL, particularly m1A, m5C, m6A, suggested regulatory mechanisms. Immune infiltration revealed correlations scores, differing by cancer type. also positively HRD several indicating association genomic instability. Enrichment analyses highlighted involvement processes oxidative phosphorylation autophagy, underscoring influence tumorigenesis. These findings establish biomarker progression regulation, warranting further exploration implications diverse types.

Language: Английский

Citations

0
A comprehensive analysis of the pyruvate kinase M1/2 (PKM) in human cancer DOI
Shuaishuai Xue, Ziyi Luo,

Yangqi Mao

et al.

Gene, Journal Year: 2024, Volume and Issue: unknown, P. 149155 - 149155

Published: Dec. 1, 2024

Language: Английский

Citations

2
Novel plasma cytokines identified and validated in children during lead exposure according to the new updated BLRV DOI Creative Commons
Xuezhong Zhang,

Lingling Sun,

Mark Lloyd Granaderos Dapar

et al.

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Dec. 5, 2024

Lead is a pervasive environmental contaminant with significant health risks, particularly to children. It known for its neurotoxic and immunotoxic effects, causing developmental, cognitive, behavioral impairments. Despite extensive research, the mechanisms of lead toxicity remain unclear. Cytokines, which are critical in immune response inflammation, have emerged as potential biomarkers toxicity. The recent Centers Disease Control Prevention (CDC) update blood reference value (BLRV) 3.5 µg/dL emphasizes need explore novel mechanisms. study involved 100 healthy children aged 1 5 years, divided into two groups based on BLRV: elevated (≥ µg/dL) low (< µg/dL). research consisted phases: discovery validation. Plasma samples were analyzed using RayBio® Human Cytokine Antibody Arrays Enzyme-linked immunosorbent assay (ELISA) cytokine levels. Ethical approval was obtained, statistical analyses included t-tests, chi-squared tests, pearson correlations, multivariate logistic regression. Protein-protein interaction (PPI), Gene Ontology (GO) enrichment, Kyoto Encyclopedia Genes Genomes (KEGG) pathway conducted roles differentially expressed proteins (DEPs). No differences age, gender, or BMI between groups, but BLRV levels significantly higher group compared group. In phase, changes expression identified, including increased IL-6, IL-8, IL-17, decreased BDNF, BMP-4, IGF-1, IL-7, IL-10, Leptin. These findings validated second phase ELISA. Significant positive correlations found IL-17. Negative observed Multivariate regression confirmed that affects these PPI networks revealed DEPs had strong interactions multiple proteins, indicating their central role GO KEGG highlighted pathways related neurotoxicity inflammatory responses, "negative regulation myotube differentiation," "neurotrophin signaling pathway," "alcoholism." This provides insights cytokines offers comprehensive analysis involved. underscore importance early detection intervention updated thresholds.

Language: Английский

Citations

0
Integrative Machine Learning Approaches to Identify and Validate Gene Biomarkers for Early Detection of Hepatocellular Carcinoma DOI Creative Commons

Mahati Munikoti Srikantamur,

Parneet Kaur,

Eckart Bindewald

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 30, 2024

Abstract Hepatocellular carcinoma (HCC) is among the leading causes of cancer-related deaths worldwide, and prognosis poor if disease detected at advanced stages. There an urgent need for early diagnostic biomarkers to facilitate timely interventions. Current methods, such as liver function tests (LFTs), alpha-fetoprotein (AFP) panels, imaging techniques like magnetic resonance (MRI) ultrasound, lack specificity HCC do not provide a comprehensive prognosis. This study proposes machine learning (ML) based approach identifying using RNA-sequencing (RNA-seq) data. We analyzed publicly available RNA-seq datasets from Gene Expression Omnibus (GEO), UCSC Xena, GEO Experiments Interactive Navigator (GREIN). In this study, we performed various feature selection methods ML with Random Forest (RF) model, achieving best performance in predicting top most significantly important genes. Bioinformatics tools, including Search Tool Retrieval Interacting Genes/Proteins (STRING), Ontology (GO), DAVID (Database Annotation, Visualization, Integrated Discovery), Human Protein Atlas (HPA), Comparative Toxicogenomics Database (CTD) were used validation. Through our analysis, identified six potential early-detection gene HCC: CDKN3, LIFR, MKI67, TOP2A, SLC5A1, VIPR1.

Language: Английский

Citations

0