MethylNet: an automated and modular deep learning approach for DNA methylation analysis

Brain cell-type shifts in Alzheimer's disease, autism, and schizophrenia interrogated using methylomics and genetics

Few neuropsychiatric disorders have replicable biomarkers, prompting high-resolution and large-scale molecular studies. However, we still lack consensus on a more foundational question: whether quantitative shifts in cell types-the functional unit of life-contribute to neuropsychiatric disorders. Leveraging advances in human brain single-cell methylomics, we deconvolve seven major cell types using bulk DNA methylation profiling across 1270 postmortem brains, including from individuals diagnosed with Alzheimer's disease, schizophrenia, and autism. We observe and replicate cell-type compositional shifts for Alzheimer's disease (endothelial cell loss), autism (increased microglia), and schizophrenia (decreased oligodendrocytes), and find age- and sex-related changes. Multiple layers of evidence indicate that endothelial cell loss contributes to Alzheimer's disease, with comparable effect size to APOE genotype among older people. Genome-wide association identified five genetic loci related to cell-type composition, involving plausible genes for the neurovascular unit (P2RX5 and TRPV3) and excitatory neurons (DPY30 and MEMO1). These results implicate specific cell-type shifts in the pathophysiology of neuropsychiatric disorders.

Biomarkers of Aging and Relevant Evaluation Techniques: A Comprehensive Review

The risk of developing chronic illnesses and disabilities is increasing with age. To predict and prevent aging, biomarkers relevant to the aging process must be identified. This paper reviews the known molecular, cellular, and physiological biomarkers of aging. Moreover, we discuss the currently available technologies for identifying these biomarkers, and their applications and potential in aging research. We hope that this review will stimulate further research and innovation in this emerging and fast-growing field.

A multi-organization epigenetic age prediction based on a channel attention perceptron networks

DNA methylation indicates the individual's aging, so-called Epigenetic clocks, which will improve the research and diagnosis of aging diseases by investigating the correlation between methylation loci and human aging. Although this discovery has inspired many researchers to develop traditional computational methods to quantify the correlation and predict the chronological age, the performance bottleneck delayed access to the practical application. Since artificial intelligence technology brought great opportunities in research, we proposed a perceptron model integrating a channel attention mechanism named PerSEClock. The model was trained on 24,516 CpG loci that can utilize the samples from all types of methylation identification platforms and tested on 15 independent datasets against seven methylation-based age prediction methods. PerSEClock demonstrated the ability to assign varying weights to different CpG loci. This feature allows the model to enhance the weight of age-related loci while reducing the weight of irrelevant loci. The method is free to use for academics at www.dnamclock.com/#/original.

Application of AI in biological age prediction

The development of anti-aging interventions requires quantitative measurement of biological age. Machine learning models, known as "aging clocks," are built by leveraging diverse aging biomarkers that vary across lifespan to predict biological age. In addition to traditional aging clocks harnessing epigenetic signatures derived from bulk samples, emerging technologies allow the biological age estimating at single-cell level to dissect cellular diversity in aging tissues. Moreover, imaging-based aging clocks are increasingly employed with the advantage of non-invasive measurement, making it suitable for large-scale human cohort studies. To fully capture the features in the ever-growing multi-modal and high-dimensional aging-related data and uncover disease associations, deep-learning based approaches, which are effective to learn complex and non-linear relationships without relying on pre-defined features, are increasingly applied. The use of big data and AI-based aging clocks has achieved high accuracy, interpretability and generalizability, guiding clinical applications to delay age-related diseases and extend healthy lifespans.

Toward Explainable Artificial Intelligence for Precision Pathology

The rapid development of precision medicine in recent years has started to challenge diagnostic pathology with respect to its ability to analyze histological images and increasingly large molecular profiling data in a quantitative, integrative, and standardized way. Artificial intelligence (AI) and, more precisely, deep learning technologies have recently demonstrated the potential to facilitate complex data analysis tasks, including clinical, histological, and molecular data for disease classification; tissue biomarker quantification; and clinical outcome prediction. This review provides a general introduction to AI and describes recent developments with a focus on applications in diagnostic pathology and beyond. We explain limitations including the black-box character of conventional AI and describe solutions to make machine learning decisions more transparent with so-called explainable AI. The purpose of the review is to foster a mutual understanding of both the biomedical and the AI side. To that end, in addition to providing an overview of the relevant foundations in pathology and machine learning, we present worked-through examples for a better practical understanding of what AI can achieve and how it should be done.

Biologically informed deep learning for explainable epigenetic clocks

Ageing is often characterised by progressive accumulation of damage, and it is one of the most important risk factors for chronic disease development. Epigenetic mechanisms including DNA methylation could functionally contribute to organismal aging, however the key functions and biological processes may govern ageing are still not understood. Although age predictors called epigenetic clocks can accurately estimate the biological age of an individual based on cellular DNA methylation, their models have limited ability to explain the prediction algorithm behind and underlying key biological processes controlling ageing. Here we present XAI-AGE, a biologically informed, explainable deep neural network model for accurate biological age prediction across multiple tissue types. We show that XAI-AGE outperforms the first-generation age predictors and achieves similar results to deep learning-based models, while opening up the possibility to infer biologically meaningful insights of the activity of pathways and other abstract biological processes directly from the model.

eXplainable Artificial Intelligence (XAI) in aging clock models

XAI is a rapidly progressing field of machine learning, aiming to unravel the predictions of complex models. XAI is especially required in sensitive applications, e.g. in health care, when diagnosis, recommendations and treatment choices might rely on the decisions made by artificial intelligence systems. AI approaches have become widely used in aging research as well, in particular, in developing biological clock models and identifying biomarkers of aging and age-related diseases. However, the potential of XAI here awaits to be fully appreciated. We discuss the application of XAI for developing the "aging clocks" and present a comprehensive analysis of the literature categorized by the focus on particular physiological systems.

ResnetAge: A Resnet-Based DNA Methylation Age Prediction Method

Aging is a significant contributing factor to degenerative diseases such as cancer. The extent of DNA methylation in human cells indicates the aging process and screening for age-related methylation sites can be used to construct epigenetic clocks. Thereby, it can be a new aging-detecting marker for clinical diagnosis and treatments. Predicting the biological age of human individuals is conducive to the study of physical aging problems. Although many researchers have developed epigenetic clock prediction methods based on traditional machine learning and even deep learning, higher prediction accuracy is still required to match the clinical applications. Here, we proposed an epigenetic clock prediction method based on a Resnet neuro networks model named ResnetAge. The model accepts 22,278 CpG sites as a sample input, supporting both the Illumina 27K and 450K identification frameworks. It was trained using 32 public datasets containing multiple tissues such as whole blood, saliva, and mouth. The Mean Absolute Error (MAE) of the training set is 1.29 years, and the Median Absolute Deviation (MAD) is 0.98 years. The Mean Absolute Error (MAE) of the validation set is 3.24 years, and the Median Absolute Deviation (MAD) is 2.3 years. Our method has higher accuracy in age prediction in comparison with other methylation-based age prediction methods.

Yearning for machine learning: applications for the classification and characterisation of senescence

Senescence is a widely appreciated tumour suppressive mechanism, which acts as a barrier to cancer development by arresting cell cycle progression in response to harmful stimuli. However, senescent cell accumulation becomes deleterious in aging and contributes to a wide range of age-related pathologies. Furthermore, senescence has beneficial roles and is associated with a growing list of normal physiological processes including wound healing and embryonic development. Therefore, the biological role of senescent cells has become increasingly nuanced and complex. The emergence of sophisticated, next-generation profiling technologies, such as single-cell RNA sequencing, has accelerated our understanding of the heterogeneity of senescence, with distinct final cell states emerging within models as well as between cell types and tissues. In order to explore data sets of increasing size and complexity, the senescence field has begun to employ machine learning (ML) methodologies to probe these intricacies. Most notably, ML has been used to aid the classification of cells as senescent, as well as to characterise the final senescence phenotypes. Here, we provide a background to the principles of ML tasks, as well as some of the most commonly used methodologies from both traditional and deep ML. We focus on the application of these within the context of senescence research, by addressing the utility of ML for the analysis of data from different laboratory technologies (microscopy, transcriptomics, proteomics, methylomics), as well as the potential within senolytic drug discovery. Together, we aim to highlight both the progress and potential for the application of ML within senescence research.

Application of deep learning in cancer epigenetics through DNA methylation analysis

DNA methylation is a fundamental epigenetic modification involved in various biological processes and diseases. Analysis of DNA methylation data at a genome-wide and high-throughput level can provide insights into diseases influenced by epigenetics, such as cancer. Recent technological advances have led to the development of high-throughput approaches, such as genome-scale profiling, that allow for computational analysis of epigenetics. Deep learning (DL) methods are essential in facilitating computational studies in epigenetics for DNA methylation analysis. In this systematic review, we assessed the various applications of DL applied to DNA methylation data or multi-omics data to discover cancer biomarkers, perform classification, imputation and survival analysis. The review first introduces state-of-the-art DL architectures and highlights their usefulness in addressing challenges related to cancer epigenetics. Finally, the review discusses potential limitations and future research directions in this field.

NCAE: data-driven representations using a deep network-coherent DNA methylation autoencoder identify robust disease and risk factor signatures

Precision medicine relies on the identification of robust disease and risk factor signatures from omics data. However, current knowledge-driven approaches may overlook novel or unexpected phenomena due to the inherent biases in biological knowledge. In this study, we present a data-driven signature discovery workflow for DNA methylation analysis utilizing network-coherent autoencoders (NCAEs) with biologically relevant latent embeddings. First, we explored the architecture space of autoencoders trained on a large-scale pan-tissue compendium (n = 75 272) of human epigenome-wide association studies. We observed the emergence of co-localized patterns in the deep autoencoder latent space representations that corresponded to biological network modules. We determined the NCAE configuration with the strongest co-localization and centrality signals in the human protein interactome. Leveraging the NCAE embeddings, we then trained interpretable deep neural networks for risk factor (aging, smoking) and disease (systemic lupus erythematosus) prediction and classification tasks. Remarkably, our NCAE embedding-based models outperformed existing predictors, revealing novel DNA methylation signatures enriched in gene sets and pathways associated with the studied condition in each case. Our data-driven biomarker discovery workflow provides a generally applicable pipeline to capture relevant risk factor and disease information. By surpassing the limitations of knowledge-driven methods, our approach enhances the understanding of complex epigenetic processes, facilitating the development of more effective diagnostic and therapeutic strategies.

Machine learning in the identification of prognostic DNA methylation biomarkers among patients with cancer: A systematic review of epigenome-wide studies

BACKGROUND

DNA methylation biomarkers have great potential in improving prognostic classification systems for patients with cancer. Machine learning (ML)-based analytic techniques might help overcome the challenges of analyzing high-dimensional data in relatively small sample sizes. This systematic review summarizes the current use of ML-based methods in epigenome-wide studies for the identification of DNA methylation signatures associated with cancer prognosis.

METHODS

We searched three electronic databases including PubMed, EMBASE, and Web of Science for articles published until 2 January 2023. ML-based methods and workflows used to identify DNA methylation signatures associated with cancer prognosis were extracted and summarized. Two authors independently assessed the methodological quality of included studies by a seven-item checklist adapted from 'A Tool to Assess Risk of Bias and Applicability of Prediction Model Studies (PROBAST)' and from the 'Reporting Recommendations for Tumor Marker Prognostic Studies (REMARK). Different ML methods and workflows used in included studies were summarized and visualized by a sunburst chart, a bubble chart, and Sankey diagrams, respectively.

RESULTS

Eighty-three studies were included in this review. Three major types of ML-based workflows were identified. 1) unsupervised clustering, 2) supervised feature selection, and 3) deep learning-based feature transformation. For the three workflows, the most frequently used ML techniques were consensus clustering, least absolute shrinkage and selection operator (LASSO), and autoencoder, respectively. The systematic review revealed that the performance of these approaches has not been adequately evaluated yet and that methodological and reporting flaws were common in the identified studies using ML techniques.

CONCLUSIONS

There is great heterogeneity in ML-based methodological strategies used by epigenome-wide studies to identify DNA methylation markers associated with cancer prognosis. In theory, most existing workflows could not handle the high multi-collinearity and potentially non-linearity interactions in epigenome-wide DNA methylation data. Benchmarking studies are needed to compare the relative performance of various approaches for specific cancer types. Adherence to relevant methodological and reporting guidelines are urgently needed.

iCpG-Pos: an accurate computational approach for identification of CpG sites using positional features on single-cell whole genome sequence data

MOTIVATION

The investigation of DNA methylation can shed light on the processes underlying human well-being and help determine overall human health. However, insufficient coverage makes it challenging to implement single-stranded DNA methylation sequencing technologies, highlighting the need for an efficient prediction model. Models are required to create an understanding of the underlying biological systems and to project single-cell (methylated) data accurately.

RESULTS

In this study, we developed positional features for predicting CpG sites. Positional characteristics of the sequence are derived using data from CpG regions and the separation between nearby CpG sites. Multiple optimized classifiers and different ensemble learning approaches are evaluated. The OPTUNA framework is used to optimize the algorithms. The CatBoost algorithm followed by the stacking algorithm outperformed existing DNA methylation identifiers.

AVAILABILITY AND IMPLEMENTATION

The data and methodologies used in this study are openly accessible to the research community. Researchers can access the positional features and algorithms used for predicting CpG site methylation patterns. To achieve superior performance, we employed the CatBoost algorithm followed by the stacking algorithm, which outperformed existing DNA methylation identifiers. The proposed iCpG-Pos approach utilizes only positional features, resulting in a substantial reduction in computational complexity compared to other known approaches for detecting CpG site methylation patterns. In conclusion, our study introduces a novel approach, iCpG-Pos, for predicting CpG site methylation patterns. By focusing on positional features, our model offers both accuracy and efficiency, making it a promising tool for advancing DNA methylation research and its applications in human health and well-being.

Artificial Intelligence-Assisted Diagnostic Cytology and Genomic Testing for Hematologic Disorders

Artificial intelligence (AI) is a rapidly evolving field of computer science that involves the development of computational programs that can mimic human intelligence. In particular, machine learning and deep learning models have enabled the identification and grouping of patterns within data, leading to the development of AI systems that have been applied in various areas of hematology, including digital pathology, alpha thalassemia patient screening, cytogenetics, immunophenotyping, and sequencing. These AI-assisted methods have shown promise in improving diagnostic accuracy and efficiency, identifying novel biomarkers, and predicting treatment outcomes. However, limitations such as limited databases, lack of validation and standardization, systematic errors, and bias prevent AI from completely replacing manual diagnosis in hematology. In addition, the processing of large amounts of patient data and personal information by AI poses potential data privacy issues, necessitating the development of regulations to evaluate AI systems and address ethical concerns in clinical AI systems. Nonetheless, with continued research and development, AI has the potential to revolutionize the field of hematology and improve patient outcomes. To fully realize this potential, however, the challenges facing AI in hematology must be addressed and overcome.

AttOmics: attention-based architecture for diagnosis and prognosis from omics data

MOTIVATION

The increasing availability of high-throughput omics data allows for considering a new medicine centered on individual patients. Precision medicine relies on exploiting these high-throughput data with machine-learning models, especially the ones based on deep-learning approaches, to improve diagnosis. Due to the high-dimensional small-sample nature of omics data, current deep-learning models end up with many parameters and have to be fitted with a limited training set. Furthermore, interactions between molecular entities inside an omics profile are not patient specific but are the same for all patients.

RESULTS

In this article, we propose AttOmics, a new deep-learning architecture based on the self-attention mechanism. First, we decompose each omics profile into a set of groups, where each group contains related features. Then, by applying the self-attention mechanism to the set of groups, we can capture the different interactions specific to a patient. The results of different experiments carried out in this article show that our model can accurately predict the phenotype of a patient with fewer parameters than deep neural networks. Visualizing the attention maps can provide new insights into the essential groups for a particular phenotype.

AVAILABILITY AND IMPLEMENTATION

The code and data are available at https://forge.ibisc.univ-evry.fr/abeaude/AttOmics. TCGA data can be downloaded from the Genomic Data Commons Data Portal.

Artificial intelligence in cancer immunotherapy: Applications in neoantigen recognition, antibody design and immunotherapy response prediction

Cancer immunotherapy is a method of controlling and eliminating tumors by reactivating the body's cancer-immunity cycle and restoring its antitumor immune response. The increased availability of data, combined with advancements in high-performance computing and innovative artificial intelligence (AI) technology, has resulted in a rise in the use of AI in oncology research. State-of-the-art AI models for functional classification and prediction in immunotherapy research are increasingly used to support laboratory-based experiments. This review offers a glimpse of the current AI applications in immunotherapy, including neoantigen recognition, antibody design, and prediction of immunotherapy response. Advancing in this direction will result in more robust predictive models for developing better targets, drugs, and treatments, and these advancements will eventually make their way into the clinical setting, pushing AI forward in the field of precision oncology.

Comparative epigenomics by machine learning approach for neuroblastoma

BACKGROUND

Neuroblastoma (NB) is the second most common pediatric solid tumor. Because the number of genetic mutations found in tumors are small, even in some patients with unfavorable NB, epigenetic variation is expected to play an important role in NB progression. DNA methylation is a major epigenetic mechanism, and its relationship with NB prognosis has been a concern. One limitation with the analysis of variation in DNA methylation is the lack of a suitable analytical model. Therefore, in this study, we performed a random forest (RF) analysis of the DNA methylome data of NB from multiple databases.

RESULTS

RF is a popular machine learning model owing to its simplicity, intuitiveness, and computational cost. RF analysis identified novel intermediate-risk patient groups with characteristic DNA methylation patterns within the low-risk group. Feature selection analysis based on probe annotation revealed that enhancer-annotated regions had strong predictive power, particularly for MYCN-amplified NBs. We developed a gene-based analytical model to identify candidate genes related to disease progression, such as PRDM8 and FAM13A-AS1. RF analysis revealed sufficient predictive power compared to other machine learning models.

CONCLUSIONS

RF is a useful tool for DNA methylome analysis in cancer epigenetic studies, and has potential to identify a novel cancer-related genes.

Investigating the prediction of CpG methylation levels from SNP genotype data to help elucidate relationships between methylation, gene expression and complex traits

As popularised by PrediXcan (and related methods), transcriptome-wide association studies (TWAS), in which gene expression is imputed from single-nucleotide polymorphism (SNP) genotypes and tested for association with a phenotype, are a popular approach for investigating the role of gene expression in complex traits. Like gene expression, DNA methylation is an important biological process and, being under genetic regulation, may be imputable from SNP genotypes. Here, we investigate prediction of CpG methylation levels from SNP genotype data to help elucidate relationships between methylation, gene expression and complex traits. We start by examining how well CpG methylation can be predicted from SNP genotypes, comparing three penalised regression approaches and examining whether changing the window size improves prediction accuracy. Although methylation at most CpG sites cannot be accurately predicted from SNP genotypes, for a subset it can be predicted well. We next apply our methylation prediction models (trained using the optimal method and window size) to carry out a methylome-wide association study (MWAS) of primary biliary cholangitis. We intersect the regions identified via MWAS with those identified via TWAS, providing insight into the interplay between CpG methylation, gene expression and disease status. We conclude that MWAS has the potential to improve understanding of biological mechanisms in complex traits.

A pan-tissue DNA-methylation epigenetic clock based on deep learning

Several age predictors based on DNA methylation, dubbed epigenetic clocks, have been created in recent years, with the vast majority based on regularized linear regression. This study explores the improvement in the performance and interpretation of epigenetic clocks using deep learning. First, we gathered 142 publicly available data sets from several human tissues to develop AltumAge, a neural network framework that is a highly accurate and precise age predictor. Compared to ElasticNet, AltumAge performs better for within-data set and cross-data set age prediction, being particularly more generalizable in older ages and new tissue types. We then used deep learning interpretation methods to learn which methylation sites contributed to the final model predictions. We observe that while most important CpG sites are linearly related to age, some highly-interacting CpG sites can influence the relevance of such relationships. Using chromatin annotations, we show that the CpG sites with the highest contribution to the model predictions were related to gene regulatory regions in the genome, including proximity to CTCF binding sites. We also found age-related KEGG pathways for genes containing these CpG sites. Lastly, we performed downstream analyses of AltumAge to explore its applicability and compare its age acceleration with Horvath’s 2013 model. We show that our neural network approach predicts higher age acceleration for tumors, for cells that exhibit age-related changes in vitro, such as immune and mitochondrial dysfunction, and for samples from patients with multiple sclerosis, type 2 diabetes, and HIV, among other conditions. Altogether, our neural network approach provides significant improvement and flexibility compared to current epigenetic clocks for both performance and model interpretability.

Epigenetic aging: Biological age prediction and informing a mechanistic theory of aging

Numerous studies have shown that epigenetic age-an individual's degree of aging based on patterns of DNA methylation-can be computed and is associated with an array of factors including diet, lifestyle, genetics, and disease. One can expect that still further associations will emerge with additional aging research, but to what end? Prediction of age was an important first step, but-in our view-the focus must shift from chasing increasingly accurate age computations to understanding the links between the epigenome and the mechanisms and physiological changes of aging. Here, we outline emerging areas of epigenetic aging research that prioritize biological understanding and clinical application. First, we survey recent progress in epigenetic clocks, which are beginning to predict not only chronological age but aging outcomes such as all-cause mortality and onset of disease, or which integrate aging signals across multiple biological processes. Second, we discuss research that exemplifies how investigation of the epigenome is building a mechanistic theory of aging and informing clinical practice. Such examples include identifying methylation sites and the genes most strongly predictive of aging-a subset of which have shown strong potential as biomarkers of neurodegenerative disease and cancer; relating epigenetic clock predictions to hallmarks of aging; and using longitudinal studies of DNA methylation to characterize human disease, resulting in the discovery of epigenetic indications of type 1 diabetes and the propensity for psychotic experiences.

Multi-task deep autoencoder to predict Alzheimer's disease progression using temporal DNA methylation data in peripheral blood

Traditional approaches for diagnosing Alzheimer's disease (AD) such as brain imaging and cerebrospinal fluid are invasive and expensive. It is desirable to develop a useful diagnostic tool by exploiting biomarkers obtained from peripheral tissues due to their noninvasive and easily accessible characteristics. However, the capacity of using DNA methylation data in peripheral blood for predicting AD progression is rarely known. It is also challenging to develop an efficient prediction model considering the complex and high-dimensional DNA methylation data in a longitudinal study. Here, we develop two multi-task deep autoencoders, which are based on the convolutional autoencoder and long short-term memory autoencoder to learn the compressed feature representation by jointly minimizing the reconstruction error and maximizing the prediction accuracy. By benchmarking on longitudinal DNA methylation data collected from the peripheral blood in Alzheimer's Disease Neuroimaging Initiative, we demonstrate that the proposed multi-task deep autoencoders outperform state-of-the-art machine learning approaches for both predicting AD progression and reconstructing the temporal DNA methylation profiles. In addition, the proposed multi-task deep autoencoders can predict AD progression accurately using only the historical DNA methylation data and the performance is further improved by including all temporal DNA methylation data. Availability:: https://github.com/lichen-lab/MTAE.

Disease classification for whole-blood DNA methylation: Meta-analysis, missing values imputation, and XAI

BACKGROUND

DNA methylation has a significant effect on gene expression and can be associated with various diseases. Meta-analysis of available DNA methylation datasets requires development of a specific workflow for joint data processing.

RESULTS

We propose a comprehensive approach of combined DNA methylation datasets to classify controls and patients. The solution includes data harmonization, construction of machine learning classification models, dimensionality reduction of models, imputation of missing values, and explanation of model predictions by explainable artificial intelligence (XAI) algorithms. We show that harmonization can improve classification accuracy by up to 20% when preprocessing methods of the training and test datasets are different. The best accuracy results were obtained with tree ensembles, reaching above 95% for Parkinson's disease. Dimensionality reduction can substantially decrease the number of features, without detriment to the classification accuracy. The best imputation methods achieve almost the same classification accuracy for data with missing values as for the original data. XAI approaches have allowed us to explain model predictions from both populational and individual perspectives.

CONCLUSIONS

We propose a methodologically valid and comprehensive approach to the classification of healthy individuals and patients with various diseases based on whole-blood DNA methylation data using Parkinson's disease and schizophrenia as examples. The proposed algorithm works better for the former pathology, characterized by a complex set of symptoms. It allows to solve data harmonization problems for meta-analysis of many different datasets, impute missing values, and build classification models of small dimensionality.

Deep Learning in High Voltage Engineering: A Literature Review

Condition monitoring of high voltage apparatus is of much importance for the maintenance of electric power systems. Whether it is detecting faults or partial discharges that take place in high voltage equipment, or detecting contamination and degradation of outdoor insulators, deep learning which is a branch of machine learning has been extensively investigated. Instead of using hand-crafted manual features as an input for the traditional machine learning algorithms, deep learning algorithms use raw data as the input where the feature extraction stage is integrated in the learning stage, resulting in a more automated process. This is the main advantage of using deep learning instead of traditional machine learning techniques. This paper presents a review of the recent literature on the application of deep learning techniques in monitoring high voltage apparatus such as GIS, transformers, cables, rotating machines, and outdoor insulators.

Systematic evaluation of cell-type deconvolution pipelines for sequencing-based bulk DNA methylomes

DNA methylation analysis by sequencing is becoming increasingly popular, yielding methylomes at single-base pair and single-molecule resolution. It has tremendous potential for cell-type heterogeneity analysis using intrinsic read-level information. Although diverse deconvolution methods were developed to infer cell-type composition based on bulk sequencing-based methylomes, systematic evaluation has not been performed yet. Here, we thoroughly benchmark six previously published methods: Bayesian epiallele detection, DXM, PRISM, csmFinder+coMethy, ClubCpG and MethylPurify, together with two array-based methods, MeDeCom and Houseman, as a comparison group. Sequencing-based deconvolution methods consist of two main steps, informative region selection and cell-type composition estimation, thus each was individually assessed. With this elaborate evaluation, we aimed to establish which method achieves the highest performance in different scenarios of synthetic bulk samples. We found that cell-type deconvolution performance is influenced by different factors depending on the number of cell types within the mixture. Finally, we propose a best-practice deconvolution strategy for sequencing data and point out limitations that need to be handled. Array-based methods—both reference-based and reference-free—generally outperformed sequencing-based methods, despite the absence of read-level information. This implies that the current sequencing-based methods still struggle with correctly identifying cell-type-specific signals and eliminating confounding methylation patterns, which needs to be handled in future studies.

Widespread redundancy in -omics profiles of cancer mutation states

BACKGROUND

In studies of cellular function in cancer, researchers are increasingly able to choose from many -omics assays as functional readouts. Choosing the correct readout for a given study can be difficult, and which layer of cellular function is most suitable to capture the relevant signal remains unclear.

RESULTS

We consider prediction of cancer mutation status (presence or absence) from functional -omics data as a representative problem that presents an opportunity to quantify and compare the ability of different -omics readouts to capture signals of dysregulation in cancer. From the TCGA Pan-Cancer Atlas that contains genetic alteration data, we focus on RNA sequencing, DNA methylation arrays, reverse phase protein arrays (RPPA), microRNA, and somatic mutational signatures as -omics readouts. Across a collection of genes recurrently mutated in cancer, RNA sequencing tends to be the most effective predictor of mutation state. We find that one or more other data types for many of the genes are approximately equally effective predictors. Performance is more variable between mutations than that between data types for the same mutation, and there is little difference between the top data types. We also find that combining data types into a single multi-omics model provides little or no improvement in predictive ability over the best individual data type.

CONCLUSIONS

Based on our results, for the design of studies focused on the functional outcomes of cancer mutations, there are often multiple -omics types that can serve as effective readouts, although gene expression seems to be a reasonable default option.

DAISM-DNNXMBD: Highly accurate cell type proportion estimation with in silico data augmentation and deep neural networks

Understanding the immune cell abundance of cancer and other disease-related tissues has an important role in guiding disease treatments. Computational cell type proportion estimation methods have been previously developed to derive such information from bulk RNA sequencing data. Unfortunately, our results show that the performance of these methods can be seriously plagued by the mismatch between training data and real-world data. To tackle this issue, we propose the DAISM-DNN^XMBD (XMBD: Xiamen Big Data, a biomedical open software initiative in the National Institute for Data Science in Health and Medicine, Xiamen University, China.) (denoted as DAISM-DNN) pipeline that trains a deep neural network (DNN) with dataset-specific training data populated from a certain amount of calibrated samples using DAISM, a novel data augmentation method with an in silico mixing strategy. The evaluation results demonstrate that the DAISM-DNN pipeline outperforms other existing methods consistently and substantially for all the cell types under evaluation in real-world datasets.

•

We propose a data augmentation method (DAISM) for DNN-based cell type deconvolution

•

DAISM-DNN enables accurate cell type deconvolution with dataset-specific training data

•

DAISM-DNN is robust to random errors in calibration samples

•

Trained DAISM-DNN model is reusable across biomedical experiments following same SOP

Computational cell type deconvolution methods were developed to understand the cellular heterogeneity in disease-related tissues from bulk RNA-seq data. Due to the presence of strong batch effects, the performance of existing methods could fluctuate greatly when applied to different datasets even with the latest development in batch normalization or platform-agnostic signature designs. To tackle this issue, we proposed a DNN-based cell abundance estimation method with dataset-specific training data populated from a certain number of calibrated samples from a target dataset using DAISM, a data augmentation method using an in silico mixing strategy. DAISM-DNN enables accurate cell type proportions prediction and is robust to random errors in the ground truth cell type proportions of calibration samples. Importantly, we showed that with strict SOPs, it is possible to create a “train once, reuse many times” DAISM-DNN model for multiple biomedical experiments without the need for retraining.

The applications of DNA methylation as a biomarker in kidney transplantation: a systematic review

Background

Although kidney transplantation improves patient survival and quality of life, long-term results are hampered by both immune- and non-immune-mediated complications. Current biomarkers of post-transplant complications, such as allograft rejection, chronic renal allograft dysfunction, and cutaneous squamous cell carcinoma, have a suboptimal predictive value. DNA methylation is an epigenetic modification that directly affects gene expression and plays an important role in processes such as ischemia/reperfusion injury, fibrosis, and alloreactive immune response. Novel techniques can quickly assess the DNA methylation status of multiple loci in different cell types, allowing a deep and interesting study of cells’ activity and function. Therefore, DNA methylation has the potential to become an important biomarker for prediction and monitoring in kidney transplantation.

Purpose of the study

The aim of this study was to evaluate the role of DNA methylation as a potential biomarker of graft survival and complications development in kidney transplantation.

Material and Methods

A systematic review of several databases has been conducted. The Newcastle–Ottawa scale and the Jadad scale have been used to assess the risk of bias for observational and randomized studies, respectively.

Results

Twenty articles reporting on DNA methylation as a biomarker for kidney transplantation were included, all using DNA methylation for prediction and monitoring. DNA methylation pattern alterations in cells isolated from different tissues, such as kidney biopsies, urine, and blood, have been associated with ischemia–reperfusion injury and chronic renal allograft dysfunction. These alterations occurred in different and specific loci. DNA methylation status has also proved to be important for immune response modulation, having a crucial role in regulatory T cell definition and activity. Research also focused on a better understanding of the role of this epigenetic modification assessment for regulatory T cells isolation and expansion for future tolerance induction-oriented therapies.

Conclusions

Studies included in this review are heterogeneous in study design, biological samples, and outcome. More coordinated investigations are needed to affirm DNA methylation as a clinically relevant biomarker important for prevention, monitoring, and intervention.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13148-022-01241-7.

Predicting drug polypharmacology from cell morphology readouts using variational autoencoder latent space arithmetic

A variational autoencoder (VAE) is a machine learning algorithm, useful for generating a compressed and interpretable latent space. These representations have been generated from various biomedical data types and can be used to produce realistic-looking simulated data. However, standard vanilla VAEs suffer from entangled and uninformative latent spaces, which can be mitigated using other types of VAEs such as β-VAE and MMD-VAE. In this project, we evaluated the ability of VAEs to learn cell morphology characteristics derived from cell images. We trained and evaluated these three VAE variants-Vanilla VAE, β-VAE, and MMD-VAE-on cell morphology readouts and explored the generative capacity of each model to predict compound polypharmacology (the interactions of a drug with more than one target) using an approach called latent space arithmetic (LSA). To test the generalizability of the strategy, we also trained these VAEs using gene expression data of the same compound perturbations and found that gene expression provides complementary information. We found that the β-VAE and MMD-VAE disentangle morphology signals and reveal a more interpretable latent space. We reliably simulated morphology and gene expression readouts from certain compounds thereby predicting cell states perturbed with compounds of known polypharmacology. Inferring cell state for specific drug mechanisms could aid researchers in developing and identifying targeted therapeutics and categorizing off-target effects in the future.

DNA Methylation Biomarkers-Based Human Age Prediction Using Machine Learning

Purpose. Age can be an important clue in uncovering the identity of persons that left biological evidence at crime scenes. With the availability of DNA methylation data, several age prediction models are developed by using statistical and machine learning methods. From epigenetic studies, it has been demonstrated that there is a close association between aging and DNA methylation. Most of the existing studies focused on healthy samples, whereas diseases may have a significant impact on human age. Therefore, in this article, an age prediction model is proposed using DNA methylation biomarkers for healthy and diseased samples. Methods. The dataset contains 454 healthy samples and 400 diseased samples from publicly available sources with age (1–89 years old). Six CpG sites are identified from this data having a high correlation with age using Pearson’s correlation coefficient. In this work, the age prediction model is developed using four different machine learning techniques, namely, Multiple Linear Regression, Support Vector Regression, Gradient Boosting Regression, and Random Forest Regression. Separate models are designed for healthy and diseased data. The data are split randomly into 80 : 20 ratios for training and testing, respectively. Results. Among all the techniques, the model designed using Random Forest Regression shows the best performance, and Gradient Boosting Regression is the second best model. In the case of healthy samples, the model achieved a MAD of 2.51 years for training data and 4.85 for testing data. Also, for diseased samples, a MAD of 3.83 years is obtained for training and 9.53 years for testing. Conclusion. These results showed that the proposed model can predict age for healthy and diseased samples.

Biophysics is reshaping our perception of the epigenome: from DNA-level to high-throughput studies

Epigenetic research holds great promise to advance our understanding of biomarkers and regulatory processes in health and disease. An increasing number of new approaches, ranging from molecular to biophysical analyses, enable identifying epigenetic changes on the level of a single gene or the whole epigenome. The aim of this review is to highlight how the field is shifting from completely molecular-biology-driven solutions to multidisciplinary strategies including more reliance on biophysical analysis tools. Biophysics not only offers technical advancements in imaging or structure analysis but also helps to explore regulatory interactions. New computational methods are also being developed to meet the demand of growing data volumes and their processing. Therefore, it is important to capture these new directions in epigenetics from a biophysical perspective and discuss current challenges as well as multiple applications of biophysical methods and tools. Specifically, we gradually introduce different biophysical research methods by first considering the DNA-level information and eventually higher-order chromatin structures. Moreover, we aim to highlight that the incorporation of bioinformatics, machine learning, and artificial intelligence into biophysical analysis allows gaining new insights into complex epigenetic processes. The gained understanding has already proven useful in translational and clinical research providing better patient stratification options or new therapeutic insights. Together, this offers a better readiness to transform bench-top experiments into industrial high-throughput applications with a possibility to employ developed methods in clinical practice and diagnostics.

Deep Learning for Human Disease Detection, Subtype Classification, and Treatment Response Prediction Using Epigenomic Data

Deep learning (DL) is a distinct class of machine learning that has achieved first-class performance in many fields of study. For epigenomics, the application of DL to assist physicians and scientists in human disease-relevant prediction tasks has been relatively unexplored until very recently. In this article, we critically review published studies that employed DL models to predict disease detection, subtype classification, and treatment responses, using epigenomic data. A comprehensive search on PubMed, Scopus, Web of Science, Google Scholar, and arXiv.org was performed following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Among 1140 initially identified publications, we included 22 articles in our review. DNA methylation and RNA-sequencing data are most frequently used to train the predictive models. The reviewed models achieved a high accuracy ranged from 88.3% to 100.0% for disease detection tasks, from 69.5% to 97.8% for subtype classification tasks, and from 80.0% to 93.0% for treatment response prediction tasks. We generated a workflow to develop a predictive model that encompasses all steps from first defining human disease-related tasks to finally evaluating model performance. DL holds promise for transforming epigenomic big data into valuable knowledge that will enhance the development of translational epigenomics.

CpG Transformer for imputation of single-cell methylomes

MOTIVATION

The adoption of current single-cell DNA methylation sequencing protocols is hindered by incomplete coverage, outlining the need for effective imputation techniques. The task of imputing single-cell (methylation) data requires models to build an understanding of underlying biological processes.

RESULTS

We adapt the transformer neural network architecture to operate on methylation matrices through combining axial attention with sliding window self-attention. The obtained CpG Transformer displays state-of-the-art performances on a wide range of scBS-seq and scRRBS-seq datasets. Furthermore, we demonstrate the interpretability of CpG Transformer and illustrate its rapid transfer learning properties, allowing practitioners to train models on new datasets with a limited computational and time budget.

AVAILABILITY AND IMPLEMENTATION

CpG Transformer is freely available at https://github.com/gdewael/cpg-transformer.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Interpretable machine learning for genomics

High-throughput technologies such as next-generation sequencing allow biologists to observe cell function with unprecedented resolution, but the resulting datasets are too large and complicated for humans to understand without the aid of advanced statistical methods. Machine learning (ML) algorithms, which are designed to automatically find patterns in data, are well suited to this task. Yet these models are often so complex as to be opaque, leaving researchers with few clues about underlying mechanisms. Interpretable machine learning (iML) is a burgeoning subdiscipline of computational statistics devoted to making the predictions of ML models more intelligible to end users. This article is a gentle and critical introduction to iML, with an emphasis on genomic applications. I define relevant concepts, motivate leading methodologies, and provide a simple typology of existing approaches. I survey recent examples of iML in genomics, demonstrating how such techniques are increasingly integrated into research workflows. I argue that iML solutions are required to realize the promise of precision medicine. However, several open challenges remain. I examine the limitations of current state-of-the-art tools and propose a number of directions for future research. While the horizon for iML in genomics is wide and bright, continued progress requires close collaboration across disciplines.

AI applications in functional genomics

We review the current applications of artificial intelligence (AI) in functional genomics. The recent explosion of AI follows the remarkable achievements made possible by "deep learning", along with a burst of "big data" that can meet its hunger. Biology is about to overthrow astronomy as the paradigmatic representative of big data producer. This has been made possible by huge advancements in the field of high throughput technologies, applied to determine how the individual components of a biological system work together to accomplish different processes. The disciplines contributing to this bulk of data are collectively known as functional genomics. They consist in studies of: i) the information contained in the DNA (genomics); ii) the modifications that DNA can reversibly undergo (epigenomics); iii) the RNA transcripts originated by a genome (transcriptomics); iv) the ensemble of chemical modifications decorating different types of RNA transcripts (epitranscriptomics); v) the products of protein-coding transcripts (proteomics); and vi) the small molecules produced from cell metabolism (metabolomics) present in an organism or system at a given time, in physiological or pathological conditions. After reviewing main applications of AI in functional genomics, we discuss important accompanying issues, including ethical, legal and economic issues and the importance of explainability.

Machine learning applications for therapeutic tasks with genomics data

Thanks to the increasing availability of genomics and other biomedical data, many machine learning algorithms have been proposed for a wide range of therapeutic discovery and development tasks. In this survey, we review the literature on machine learning applications for genomics through the lens of therapeutic development. We investigate the interplay among genomics, compounds, proteins, electronic health records, cellular images, and clinical texts. We identify 22 machine learning in genomics applications that span the whole therapeutics pipeline, from discovering novel targets, personalizing medicine, developing gene-editing tools, all the way to facilitating clinical trials and post-market studies. We also pinpoint seven key challenges in this field with potentials for expansion and impact. This survey examines recent research at the intersection of machine learning, genomics, and therapeutic development.

Deep learning in cancer diagnosis, prognosis and treatment selection

Deep learning is a subdiscipline of artificial intelligence that uses a machine learning technique called artificial neural networks to extract patterns and make predictions from large data sets. The increasing adoption of deep learning across healthcare domains together with the availability of highly characterised cancer datasets has accelerated research into the utility of deep learning in the analysis of the complex biology of cancer. While early results are promising, this is a rapidly evolving field with new knowledge emerging in both cancer biology and deep learning. In this review, we provide an overview of emerging deep learning techniques and how they are being applied to oncology. We focus on the deep learning applications for omics data types, including genomic, methylation and transcriptomic data, as well as histopathology-based genomic inference, and provide perspectives on how the different data types can be integrated to develop decision support tools. We provide specific examples of how deep learning may be applied in cancer diagnosis, prognosis and treatment management. We also assess the current limitations and challenges for the application of deep learning in precision oncology, including the lack of phenotypically rich data and the need for more explainable deep learning models. Finally, we conclude with a discussion of how current obstacles can be overcome to enable future clinical utilisation of deep learning.

Epigenetic age prediction

Advanced age is the main common risk factor for cancer, cardiovascular disease and neurodegeneration. Yet, more is known about the molecular basis of any of these groups of diseases than the changes that accompany ageing itself. Progress in molecular ageing research was slow because the tools predicting whether someone aged slowly or fast (biological age) were unreliable. To understand ageing as a risk factor for disease and to develop interventions, the molecular ageing field needed a quantitative measure; a clock for biological age. Over the past decade, a number of age predictors utilising DNA methylation have been developed, referred to as epigenetic clocks. While they appear to estimate biological age, it remains unclear whether the methylation changes used to train the clocks are a reflection of other underlying cellular or molecular processes, or whether methylation itself is involved in the ageing process. The precise aspects of ageing that the epigenetic clocks capture remain hidden and seem to vary between predictors. Nonetheless, the use of epigenetic clocks has opened the door towards studying biological ageing quantitatively, and new clocks and applications, such as forensics, appear frequently. In this review, we will discuss the range of epigenetic clocks available, their strengths and weaknesses, and their applicability to various scientific queries.

MethylSPWNet and MethylCapsNet: Biologically Motivated Organization of DNAm Neural Networks, Inspired by Capsule Networks

DNA methylation (DNAm) alterations have been heavily implicated in carcinogenesis and the pathophysiology of diseases through upstream regulation of gene expression. DNAm deep-learning approaches are able to capture features associated with aging, cell type, and disease progression, but lack incorporation of prior biological knowledge. Here, we present modular, user-friendly deep-learning methodology and software, MethylCapsNet and MethylSPWNet, that group CpGs into biologically relevant capsules-such as gene promoter context, CpG island relationship, or user-defined groupings-and relate them to diagnostic and prognostic outcomes. We demonstrate these models' utility on 3,897 individuals in the classification of central nervous system (CNS) tumors. MethylCapsNet and MethylSPWNet provide an opportunity to increase DNAm deep-learning analyses' interpretability by enabling a flexible organization of DNAm data into biologically relevant capsules.

Detection of Cell Types Contributing to Cancer From Circulating, Cell-Free Methylated DNA

Detection of cellular changes in tissue biopsies has been the basis for cancer diagnostics. However, tissue biopsies are invasive and limited by inaccuracies due to sampling locations, restricted sampling frequency, and poor representation of tissue heterogeneity. Liquid biopsies are emerging as a complementary approach to traditional tissue biopsies to detect dynamic changes in specific cell populations. Cell-free DNA (cfDNA) fragments released into the circulation from dying cells can be traced back to the tissues and cell types they originated from using DNA methylation, an epigenetic regulatory mechanism that is highly cell-type specific. Decoding changes in the cellular origins of cfDNA over time can reveal altered host tissue homeostasis due to local cancer invasion and metastatic spread to distant organs as well as treatment responses. In addition to host-derived cfDNA, changes in cancer cells can be detected from cell-free, circulating tumor DNA (ctDNA) by monitoring DNA mutations carried by cancer cells. Here, we will discuss computational approaches to identify and validate robust biomarkers of changed tissue homeostasis using cell-free, methylated DNA in the circulation. We highlight studies performing genome-wide profiling of cfDNA methylation and those that combine genetic and epigenetic markers to further identify cell-type specific signatures. Finally, we discuss opportunities and current limitations of these approaches for implementation in clinical oncology.

DeepRMethylSite: a deep learning based approach for prediction of arginine methylation sites in proteins

Methylation, which is one of the most prominent post-translational modifications on proteins, regulates many important cellular functions. Though several model-based methylation site predictors have been reported, all existing methods employ machine learning strategies, such as support vector machines and random forest, to predict sites of methylation based on a set of "hand-selected" features. As a consequence, the subsequent models may be biased toward one set of features. Moreover, due to the large number of features, model development can often be computationally expensive. In this paper, we propose an alternative approach based on deep learning to predict arginine methylation sites. Our model, which we termed DeepRMethylSite, is computationally less expensive than traditional feature-based methods while eliminating potential biases that can arise through features selection. Based on independent testing on our dataset, DeepRMethylSite achieved efficiency scores of 68%, 82% and 0.51 with respect to sensitivity (SN), specificity (SP) and Matthew's correlation coefficient (MCC), respectively. Importantly, in side-by-side comparisons with other state-of-the-art methylation site predictors, our method performs on par or better in all scoring metrics tested.

Verifying explainability of a deep learning tissue classifier trained on RNA-seq data

For complex machine learning (ML) algorithms to gain widespread acceptance in decision making, we must be able to identify the features driving the predictions. Explainability models allow transparency of ML algorithms, however their reliability within high-dimensional data is unclear. To test the reliability of the explainability model SHapley Additive exPlanations (SHAP), we developed a convolutional neural network to predict tissue classification from Genotype-Tissue Expression (GTEx) RNA-seq data representing 16,651 samples from 47 tissues. Our classifier achieved an average F1 score of 96.1% on held-out GTEx samples. Using SHAP values, we identified the 2423 most discriminatory genes, of which 98.6% were also identified by differential expression analysis across all tissues. The SHAP genes reflected expected biological processes involved in tissue differentiation and function. Moreover, SHAP genes clustered tissue types with superior performance when compared to all genes, genes detected by differential expression analysis, or random genes. We demonstrate the utility and reliability of SHAP to explain a deep learning model and highlight the strengths of applying ML to transcriptome data.

DNA Methylation Associated With Diabetic Kidney Disease in Blood-Derived DNA

A subset of individuals with type 1 diabetes will develop diabetic kidney disease (DKD). DKD is heritable and large-scale genome-wide association studies have begun to identify genetic factors that influence DKD. Complementary to genetic factors, we know that a person’s epigenetic profile is also altered with DKD. This study reports analysis of DNA methylation, a major epigenetic feature, evaluating methylome-wide loci for association with DKD. Unique features (n = 485,577; 482,421 CpG probes) were evaluated in blood-derived DNA from carefully phenotyped White European individuals diagnosed with type 1 diabetes with (cases) or without (controls) DKD (n = 677 samples). Explicitly, 150 cases were compared to 100 controls using the 450K array, with subsequent analysis using data previously generated for a further 96 cases and 96 controls on the 27K array, and de novo methylation data generated for replication in 139 cases and 96 controls. Following stringent quality control, raw data were quantile normalized and beta values calculated to reflect the methylation status at each site. The difference in methylation status was evaluated between cases and controls; resultant P-values for array-based data were adjusted for multiple testing. Genes with significantly increased (hypermethylated) and/or decreased (hypomethylated) levels of DNA methylation were considered for biological relevance by functional enrichment analysis using KEGG pathways. Twenty-two loci demonstrated statistically significant fold changes associated with DKD and additional support for these associated loci was sought using independent samples derived from patients recruited with similar inclusion criteria. Markers associated with CCNL1 and ZNF187 genes are supported as differentially regulated loci (P < 10^–8), with evidence also presented for AFF3, which has been identified from a meta-analysis and subsequent replication of genome-wide association studies. Further supporting evidence for differential gene expression in CCNL1 and ZNF187 is presented from kidney biopsy and blood-derived RNA in people with and without kidney disease from NephroSeq. Evidence confirming that methylation sites influence the development of DKD may aid risk prediction tools and stimulate research to identify epigenomic therapies which might be clinically useful for this disease.

Don't dismiss logistic regression: the case for sensible extraction of interactions in the era of machine learning

BACKGROUND

Machine learning approaches have become increasingly popular modeling techniques, relying on data-driven heuristics to arrive at its solutions. Recent comparisons between these algorithms and traditional statistical modeling techniques have largely ignored the superiority gained by the former approaches due to involvement of model-building search algorithms. This has led to alignment of statistical and machine learning approaches with different types of problems and the under-development of procedures that combine their attributes. In this context, we hoped to understand the domains of applicability for each approach and to identify areas where a marriage between the two approaches is warranted. We then sought to develop a hybrid statistical-machine learning procedure with the best attributes of each.

METHODS

We present three simple examples to illustrate when to use each modeling approach and posit a general framework for combining them into an enhanced logistic regression model building procedure that aids interpretation. We study 556 benchmark machine learning datasets to uncover when machine learning techniques outperformed rudimentary logistic regression models and so are potentially well-equipped to enhance them. We illustrate a software package, InteractionTransformer, which embeds logistic regression with advanced model building capacity by using machine learning algorithms to extract candidate interaction features from a random forest model for inclusion in the model. Finally, we apply our enhanced logistic regression analysis to two real-word biomedical examples, one where predictors vary linearly with the outcome and another with extensive second-order interactions.

RESULTS

Preliminary statistical analysis demonstrated that across 556 benchmark datasets, the random forest approach significantly outperformed the logistic regression approach. We found a statistically significant increase in predictive performance when using hybrid procedures and greater clarity in the association with the outcome of terms acquired compared to directly interpreting the random forest output.

CONCLUSIONS

When a random forest model is closer to the true model, hybrid statistical-machine learning procedures can substantially enhance the performance of statistical procedures in an automated manner while preserving easy interpretation of the results. Such hybrid methods may help facilitate widespread adoption of machine learning techniques in the biomedical setting.