Machine Learning Approaches to Classify Primary and Metastatic Cancers Using Tissue of Origin-Based DNA Methylation Profiles

Modhukur, Vijayachitra; Sharma, Shakshi; Mondal, Mainak; Lawarde, Ankita; Kask, Keiu; Sharma, Rajesh; Salumets, Andres

doi:10.3390/cancers13153768

Cited by 24 publications

(14 citation statements)

References 42 publications

(57 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The results are compared with the recent work by Zheng et al [ 17 ] and Modhukur et al [ 18 ] for the compatible cancer types from primary sources.…”

Section: Resultsmentioning

confidence: 90%

“…The results are compared with the recent work by Zheng et al [17] and Modhukur et al [18] for the compatible cancer types from primary sources. The MCC values of the pan-classification DNN models were +0.981 and +0.984 for the 27k DNA methylation profile and the 450k DNA methylation profiles, respectively.…”

Section: Dnn Pan-classificationmentioning

confidence: 89%

“…In 2021, Modhukur et al [ 18 ] used different machine learning approaches to classify primary and metastasised cancers using DNA methylation samples retrieved from TCGA and other sources. They applied Support Vector Machines (SVM), Extreme Gradient Boosting, Naive Bayes (NB), and Random Forest (RF) approaches to classify the cancer types.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A hybrid metaheuristic-deep learning technique for the pan-classification of cancer based on DNA methylation

2022

View full text Add to dashboard Cite

Background DNA 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). Results In 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. Conclusions This 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.

show abstract

“…The results are compared with the recent work by Zheng et al [ 17 ] and Modhukur et al [ 18 ] for the compatible cancer types from primary sources.…”

Section: Resultsmentioning

confidence: 90%

Section: Dnn Pan-classificationmentioning

confidence: 89%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A hybrid metaheuristic-deep learning technique for the pan-classification of cancer based on DNA methylation

2022

View full text Add to dashboard Cite

show abstract

“…The remarkable tissue specificity of epigenetic characteristics also been exploited for TOO prediction, which is featured by tissue origin classification based on DNA methylation patterns [ 28 , 29 ]. In the EUICUP study, DNA methylation profiling predicted TOO in 87% (188/216) patients with CUP [ 15 ].…”

Section: Discussionmentioning

confidence: 99%

A molecular approach integrating genomic and DNA methylation profiling for tissue of origin identification in lung-specific cancer of unknown primary

Chen

Zhang

et al. 2022

J Transl Med

View full text Add to dashboard Cite

Background Determining the tissue of origin (TOO) is essential for managing cancer of unknown primary (CUP). In this study, we evaluated the concordance between genome profiling and DNA methylation analysis in determining TOO for lung-specific CUP and assessed their performance by comparing the clinical responses and survival outcomes of patients predicted with multiple primary or with metastatic cancer. Methods We started by retrospectively screening for CUP patients who presented with both intra- and extrathoracic tumors. Tumor samples from included patients were analyzed with targeted sequencing with a 520-gene panel and targeted bisulfite sequencing. TOO inferences were made in parallel via an algorithm using genome profiles and time interval between tumors and via machine learning-based classification of DNA methylation profiles. Results Four hundred patients were screened retrospectively. Excluding patients definitively diagnosed with conventional diagnostic work-up or without available samples, 16 CUP patients were included. Both molecular approaches alone enabled inference of clonality for all analyzed patients. Genome profile enabled TOO inference for 43.8% (7/16) patients, and the percentage rose to 68.8% (11/16) after considering inter-tumor time lag. On the other hand, DNA methylation analysis was conclusive for TOO prediction for 100% (14/14) patients with available samples. The two approaches gave 100% (9/9) concordant inferences regarding clonality and TOO identity. Moreover, patients predicted with metastatic disease showed significantly shorter overall survival than those with multiple primary tumors. Conclusions Genome and DNA methylation profiling have shown promise as individual analysis for TOO identification. This study demonstrated the feasibility of incorporating the two methods and proposes an integrative scheme to facilitate diagnosing and treating lung-specific CUPs.

show abstract

“…The whole-genome DNA methylation profile (methylome) of a tumor is the result of both somatically acquired changes and of features reflecting the cell of origin [36]. Thus, methylome analysis has been successful in both subclassifying tumors previously considered homogeneous diseases and in tracing the origin of undifferentiated metastases of cancers of unknown primary [37][38][39].…”

Section: Whole-genome Methylation Profiling (Methylome) Of Glioblastomasmentioning

confidence: 99%

MGMT and Whole-Genome DNA Methylation Impacts on Diagnosis, Prognosis and Therapy of Glioblastoma Multiforme

Monica

Cuomo

Buonaiuto

et al. 2022

IJMS

View full text Add to dashboard Cite

Epigenetic changes in DNA methylation contribute to the development of many diseases, including cancer. In glioblastoma multiforme, the most prevalent primary brain cancer and an incurable tumor with a median survival time of 15 months, a single epigenetic modification, the methylation of the O6-Methylguanine-DNA Methyltransferase (MGMT) gene, is a valid biomarker for predicting response to therapy with alkylating agents and also, independently, prognosis. More recently, the progress from single gene to whole-genome analysis of DNA methylation has allowed a better subclassification of glioblastomas. Here, we review the clinically relevant information that can be obtained by studying MGMT gene and whole-genome DNA methylation changes in glioblastomas, also highlighting benefits, including those of liquid biopsy, and pitfalls of the different detection methods. Finally, we discuss how changes in DNA methylation, especially in glioblastomas bearing mutations in the Isocitrate Dehydrogenase (IDH) 1 and 2 genes, can be exploited as targets for tailoring therapy.

show abstract

Machine Learning Approaches to Classify Primary and Metastatic Cancers Using Tissue of Origin-Based DNA Methylation Profiles

Cited by 24 publications

References 42 publications

A hybrid metaheuristic-deep learning technique for the pan-classification of cancer based on DNA methylation

A hybrid metaheuristic-deep learning technique for the pan-classification of cancer based on DNA methylation

A molecular approach integrating genomic and DNA methylation profiling for tissue of origin identification in lung-specific cancer of unknown primary

MGMT and Whole-Genome DNA Methylation Impacts on Diagnosis, Prognosis and Therapy of Glioblastoma Multiforme

Contact Info

Product

Resources

About