Dataset Information

Hierarchical classification-based pan-cancer methylation analysis to classify primary cancer.

ABSTRACT: Hierarchical classification offers a more specific categorization of data and breaks down large classification problems into subproblems, providing improved prediction accuracy and predictive power for undefined categories, while also mitigating the impact of poor-quality data. Despite these advantages, its application in predicting primary cancer is rare. To leverage the similarity of cancers and the specificity of methylation patterns among them, we developed the Cancer Hierarchy Classification Tool (CHCT) using the idea of hierarchical classification, with methylation data from 30 cancer types and 8239 methylome samples downloaded from publicly available databases (The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO)). We used unsupervised clustering to divide the classification subproblems and screened differentially methylated sites using Analysis of variance (ANOVA) test, Tukey-kramer test, and Boruta algorithms to construct models for each classifier module. After validation, CHCT accurately classified 1568 out of 1660 cases in the test set, with an average accuracy of 94.46%. We further curated an independent validation cohort of 677 cancer samples from GEO and assigned a diagnosis using CHCT, which showed high diagnostic potential with generally high accuracies (an average accuracy of 91.40%). Moreover, CHCT demonstrates predictive capability for additional cancer types beyond its original classifier scope as demonstrated in the medulloblastoma and pituitary tumor datasets. In summary, CHCT can hierarchically classify primary cancer by methylation profile, by splitting a large-scale classification of 30 cancer types into ten smaller classification problems. These results indicate that cancer hierarchical classification has the potential to be an accurate and robust cancer classification method.

SUBMITTER: Yang Y

PROVIDER: S-EPMC10709847 | biostudies-literature | 2023 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Hierarchical classification-based pan-cancer methylation analysis to classify primary cancer.

Yang Youpeng Y Zeng Qiuhong Q Liu Gaotong G Zheng Shiyao S Luo Tianyang T Guo Yibin Y Tang Jia J Huang Yi Y

BMC bioinformatics 20231208 1

Hierarchical classification offers a more specific categorization of data and breaks down large classification problems into subproblems, providing improved prediction accuracy and predictive power for undefined categories, while also mitigating the impact of poor-quality data. Despite these advantages, its application in predicting primary cancer is rare. To leverage the similarity of cancers and the specificity of methylation patterns among them, we developed the Cancer Hierarchy Classificatio ...[more]

PMID: 38066424

Similar Datasets

Project description:BackgroundDNA Methylation is one of the most important epigenetic processes that are crucial to regulating the functioning of the human genome without altering the DNA sequence. DNA Methylation data for cancer patients are becoming more accessible than ever, which is attributed to newer DNA sequencing technologies, notably, the relatively low-cost DNA microarray technology by Illumina Infinium. This technology makes it possible to study DNA methylation at hundreds of thousands of different loci. Currently, most of the research found in the literature focuses on the discovery of DNA methylation markers for specific cancer types. A relatively small number of studies have attempted to find unified DNA methylation biomarkers that can diagnose different types of cancer (pan-cancer classification).ResultsIn this study, the aim is to conduct a pan-classification of cancer disease. We retrieved individual data for different types of cancer patients from The Cancer Genome Atlas (TCGA) portal. We selected data for many cancer types: Breast Cancer (BRCA), Ovary Cancer (OV), Stomach Cancer (STOMACH), Colon Cancer (COAD), Kidney Cancer (KIRC), Liver Cancer (LIHC), Lung Cancer (LUSC), Prostate Cancer (PRAD) and Thyroid cancer (THCA). The data was pre-processed and later used to build the required dataset. The system that we developed consists of two main stages. The purpose of the first stage is to perform feature selection and, therefore, decrease the dimensionality of the DNA methylation loci (features). This is accomplished using an unsupervised metaheuristic technique. As for the second stage, we used supervised machine learning and developed deep neural network (DNN) models to help classify the samples' malignancy status and cancer type. Experimental results showed that compared to recently published methods, our proposed system achieved better classification results in terms of recall, and similar and higher results in terms of precision and accuracy. The proposed system also achieved an excellent receiver operating characteristic area under the curve (ROC AUC) values varying from 0.85 to 0.89.ConclusionsThis research presented an effective new approach to classify different cancer types based on DNA methylation data retrieved from TCGA. The performance of the proposed system was compared to recently published works, using different performance metrics. It provided better results, confirming the effectiveness of the proposed method for classifying different cancer types based on DNA methylation data.

Project description:The metastatic cancer of unknown primary (CUP) sites remains a leading cause of cancer death with few therapeutic options. The aberrant DNA methylation (DNAm) is the most important risk factor for cancer, which has certain tissue specificity. However, how DNAm alterations in tumors differ among the regulatory network of multi-omics remains largely unexplored. Therefore, there is room for improvement in our accuracy in the prediction of tumor origin sites and a need for better understanding of the underlying mechanisms. In our study, an integrative analysis based on multi-omics data and molecular regulatory network uncovered genome-wide methylation mechanism and identified 23 epi-driver genes. Apart from the promoter region, we also found that the aberrant methylation within the gene body or intergenic region was significantly associated with gene expression. Significant enrichment analysis of the epi-driver genes indicated that these genes were highly related to cellular mechanisms of tumorigenesis, including T-cell differentiation, cell proliferation, and signal transduction. Based on the ensemble algorithm, six CpG sites located in five epi-driver genes were selected to construct a tissue-specific classifier with a better accuracy (>95%) using TCGA datasets. In the independent datasets and the metastatic cancer datasets from GEO, the accuracy of distinguishing tumor subtypes or original sites was more than 90%, showing better robustness and stability. In summary, the integration analysis of large-scale omics data revealed complex regulation of DNAm across various cancer types and identified the epi-driver genes participating in tumorigenesis. Based on the aberrant methylation status located in epi-driver genes, a classifier that provided the highest accuracy in tracing back to the primary sites of metastatic cancer was established. Our study provides a comprehensive and multi-omics view of DNAm-associated changes across cancer types and has potential for clinical application.

Project description:BackgroundDNA methylation is a key epigenetic regulator contributing to cancer development. To understand the role of DNA methylation in tumorigenesis, it is important to investigate and compare differential methylation (DM) patterns between normal and case samples across different cancer types. However, current pan-cancer analyses call DM separately for each cancer, which suffers from lower statistical power and fails to provide a comprehensive view for patterns across cancers.MethodsIn this work, we propose a rigorous statistical model, PanDM, to jointly characterize DM patterns across diverse cancer types. PanDM uses the hidden correlations in the combined dataset to improve statistical power through joint modeling. PanDM takes summary statistics from separate analyses as input and performs methylation site clustering, differential methylation detection, and pan-cancer pattern discovery. We demonstrate the favorable performance of PanDM using simulation data. We apply our model to 12 cancer methylome data collected from The Cancer Genome Atlas (TCGA) project. We further conduct ontology- and pathway-enrichment analyses to gain new biological insights into the pan-cancer DM patterns learned by PanDM.ResultsPanDM outperforms two types of separate analyses in the power of DM calling in the simulation study. Application of PanDM to TCGA data reveals 37 pan-cancer DM patterns in the 12 cancer methylomes, including both common and cancer-type-specific patterns. These 37 patterns are in turn used to group cancer types. Functional ontology and biological pathways enriched in the non-common patterns not only underpin the cancer-type-specific etiology and pathogenesis but also unveil the common environmental risk factors shared by multiple cancer types. Moreover, we also identify PanDM-specific DM CpG sites that the common strategy fails to detect.ConclusionsPanDM is a powerful tool that provides a systematic way to investigate aberrant methylation patterns across multiple cancer types. Results from real data analyses suggest a novel angle for us to understand the common and specific DM patterns in different cancers. Moreover, as PanDM works on the summary statistics for each cancer type, the same framework can in principle be applied to pan-cancer analyses of other functional genomic profiles. We implement PanDM as an R package, which is freely available at http://www.sta.cuhk.edu.hk/YWei/PanDM.html .

Dataset Information

Hierarchical classification-based pan-cancer methylation analysis to classify primary cancer.

Publications

Hierarchical classification-based pan-cancer methylation analysis to classify primary cancer.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets