Deciphering the Methylation Landscape in Breast Cancer: Diagnostic and Prognostic Biosignatures through Automated Machine Learning

Abstract

DNA methylation plays an important role in breast cancer (BrCa) pathogenesis and could contribute to driving its personalized management. We performed a complete bioinformatic analysis in BrCa whole methylome datasets, analyzed using the Illumina methylation 450 bead-chip array. Differential methylation analysis vs. clinical end-points resulted in 11,176 to 27,786 differentially methylated genes (DMGs). Innovative automated machine learning (AutoML) was employed to construct signatures with translational value. Three highly performing and low-feature-number signatures were built: (1) A 5-gene signature discriminating BrCa patients from healthy individuals (area under the curve (AUC): 0.994 (0.982–1.000)). (2) A 3-gene signature identifying BrCa metastatic disease (AUC: 0.986 (0.921–1.000)). (3) Six equivalent 5-gene signatures diagnosing early disease (AUC: 0.973 (0.920–1.000)). Validation in independent patient groups verified performance. Bioinformatic tools for functional analysis and protein interaction prediction were also employed. All protein encoding features included in the signatures were associated with BrCa-related pathways. Functional analysis of DMGs highlighted the regulation of transcription as the main biological process, the nucleus as the main cellular component and transcription factor activity and sequence-specific DNA binding as the main molecular functions. Overall, three high-performance diagnostic/prognostic signatures were built and are readily available for improving BrCa precision management upon prospective clinical validation. Revisiting archived methylomes through novel bioinformatic approaches revealed significant clarifying knowledge for the contribution of gene methylation events in breast carcinogenesis.

Methylation Status of Corticotropin-Releasing Factor (CRF) Receptor Genes in Colorectal Cancer

“…JADBio has previously been successfully used to produce signatures for clinical applications such as development of classifiers for metastatic BrCa based on novel ccfDNA methylation patterns [ 16 ], identification of risk of lung cancer in smokers [ 72 ] or suicide prediction among depressive patients [ 73 ]. Recently, by revisiting publicly available omics datasets via JADBio, we were able to deliver accurate highly-performing blood-based predictive biosignatures in Alzheimer’s disease [ 74 ] and in breast cancer [ 75 ].…”

ENPP2 Methylation in Health and Cancer

“…Epigenetic regulation of ENPP2 has been previously reported [ 13 ]. DNA methylation, a well-studied epigenetic mechanism, can regulate gene expression [ 14 ], and aberrant gene-specific methylation has been correlated with many pathologies, such as cancer [ 15 , 16 , 17 , 18 , 19 , 20 ]. However, data on the methylation profile of ENPP2 in health and pathology are fragmented.…”

Automated machine learning optimizes and accelerates predictive modeling from COVID-19 high throughput datasets

“…All datasets used in the paper are publicly available. Original data for datasets analyzed can be found in the following publications: Lieberman et al 28 ; Mick et al 27 and Shen et al 9 .…”

“…JADBio has been extensively validated on 360 omics datasets 20 demonstrating correct estimates of performance for the produced models on the training data population. By employing this AutoML approach, we have recently produced accurate and validated biosignatures by revisiting high throughput data for Alzheimer's diagnosis and cancer prognosis, as well as modeling in various other domains [20][21][22][23][24][25][26][27] . Moreover, a COVID-19 model that estimates the probability of viral mutations to produce severe infections has been generated and is currently available in the form of an online resource 28 .…”