LogLoss-BERAF: An ensemble-based machine learning model for constructing highly accurate diagnostic sets of methylation sites accounting for heterogeneity in prostate cancer

Abstract

Although modern methods of whole genome DNA methylation analysis have a wide range of applications, they are not suitable for clinical diagnostics due to their high cost and complexity and due to the large amount of sample DNA required for the analysis. Therefore, it is crucial to be able to identify a relatively small number of methylation sites that provide high precision and sensitivity for the diagnosis of pathological states. We propose an algorithm for constructing limited subsamples from high-dimensional data to form diagnostic panels. We have developed a tool that utilizes different methods of selection to find an optimal, minimum necessary combination of factors using cross-entropy loss metrics (LogLoss) to identify a subset of methylation sites. We show that the algorithm can work effectively with different genome methylation patterns using ensemble-based machine learning methods. Algorithm efficiency, precision and robustness were evaluated using five genome-wide DNA methylation datasets (totaling 626 samples), and each dataset was classified into tumor and non-tumor samples. The algorithm produced an AUC of 0.97 (95% CI: 0.94–0.99, 9 sites) for prostate adenocarcinoma and an AUC of 1.0 (from 2 to 6 sites) for urothelial bladder carcinoma, two types of kidney carcinoma and colorectal carcinoma. For prostate adenocarcinoma we showed that identified differential variability methylation patterns distinguish cluster of samples with higher recurrence rate (hazard ratio for recurrence = 0.48, 95% CI: 0.05–0.92; log-rank test, p-value < 0.03). We also identified several clusters of correlated interchangeable methylation sites that can be used for the elaboration of biological interpretation of the resulting models and for further selection of the sites most suitable for designing diagnostic panels. LogLoss-BERAF is implemented as a standalone python code and open-source code is freely available from https://github.com/bioinformatics-IBCH/logloss-beraf along with the models described in this article.

TP63-TRIM29 axis regulate enhancer methylation and chromosomal instability in prostate cancer

“…Also, the level of methylation of TP63-dependent CpG sites can be a potential biomarker of the level of TP63 expression and consequently the level of TRIM29 expression, which strongly correlates with TP63 expression. This opens up an opportunity to detect the level of expression of these two genes by the level of methylation of CpG sites of extracellular DNA, which can be detected in urine or blood plasma (Babalyan et al 2018).…”

“…Also, the level of methylation of TP63-dependent CpG sites can be a potential biomarker of TP63 and TRIM29 levels. If so, TP63 and TRIM29 levels can be indirectly evaluated by quantification of CpG methylation in extracellular DNA from urine or blood plasma (Babalyan et al 2018).…”

Three Layer Super Learner Ensemble with Hyperparameter Optimization to Improve the Performance of Machine Learning Model

“…The LLWSL is represented using a pseudocode as shown in below Figure 1. the predicted model with some considerable level of performance though the hyperparameter tuning has not been done [3]. After a comparison of the weights of optimal base learners were selected.…”

Machine Learning and Deep Learning based AI Tools for Development of Diagnostic Tools

“…Support vector machine algorithm was used with image data for building a diagnostic tool for diagnosis of cancer ( Sweilam et al, 2010 ), Alzheimer’s disease detection with MRI scan image data ( Khedher et al, 2015 ), and a diagnostic model for predicting stroke mortality at discharge with Image data ( Ho et al, 2014 ). Similarly, a diagnostic model for heterogeneity in prostate cancer was built using DNA methylation data using an ensemble algorithm ( Babalyan et al, 2018 ).…”