A Computational Approach to Identification of Candidate Biomarkers in High-Dimensional Molecular Data

Abstract

Complex high-dimensional datasets that are challenging to analyze are frequently produced through ‘-omics’ profiling. Typically, these datasets contain more genomic features than samples, limiting the use of multivariable statistical and machine learning-based approaches to analysis. Therefore, effective alternative approaches are urgently needed to identify features-of-interest in ‘-omics’ data. In this study, we present the molecular feature selection tool, a novel, ensemble-based, feature selection application for identifying candidate biomarkers in ‘-omics’ data. As proof-of-principle, we applied the molecular feature selection tool to identify a small set of immune-related genes as potential biomarkers of three prostate adenocarcinoma subtypes. Furthermore, we tested the selected genes in a model to classify the three subtypes and compared the results to models built using all genes and all differentially expressed genes. Genes identified with the molecular feature selection tool performed better than the other models in this study in all comparison metrics: accuracy, precision, recall, and F1-score using a significantly smaller set of genes. In addition, we developed a simple graphical user interface for the molecular feature selection tool, which is available for free download. This user-friendly interface is a valuable tool for the identification of potential biomarkers in gene expression datasets and is an asset for biomarker discovery studies.

DNA polymerase ε and δ variants drive mutagenesis in polypurine tracts in human tumors

“…To select mutational motifs most predictive of the POLE driver status, feature selection was performed using MFeaST . 85 Classification model development was carried out with the scikit-learn python package (v. 1.1.1) and the MATLAB Classification Learner App with 5-fold cross-validation and default tuning hyper-parameters. To identify the most accurate classification model, 91 several machine learning classifiers were applied, namely linear discriminant, support vector machine, gradient boosting, AdaBoostClassifier, Bagging classifier, KNeighborsClassifier, Decision Tree, Random Forest, Gaussian Naive Bayes.…”

A user-driven machine learning approach for RNA-based sample discrimination and hierarchical classification

“…To perform feature selection, we recommend using an ensemble feature selection tool, MFeaST . 10 We have successfully applied this tool in many omics studies. 1 , 11 , 12 , 14 , 15 , 16 , 17 , 47 , 48 , 49 For download instructions, see before you begin – software installation and directory set-up .…”

“…To overcome this issue, we use a more generalizable ensemble feature selection approach that uses multiple feature selection algorithms. The Molecular Feature Selection Tool (MFeaST) 10 ranks all available features based on their combined score from different types of feature selection algorithms. Compared to other ensemble approaches, which are limited to a subset of feature selection algorithms, MFeaST uses filter, wrapper, and embedded techniques.…”

“… 12 Table S3 Software and algorithms MATLAB ver R2020 or later MathWorks https://www.mathworks.com/products/matlab.html Molecular Feature Selection Tool (MFeaST) Gerolami et al. 10 https://www.renwicklab.com/molecular-feast/ Custom MATLAB scripts and functions This paper and GitHub page Supplementary Materials of this protocol, https://github.com/Renwick-Lab/RNA-ML-Sample-Hierarchical-Classification ( https://doi.org/10.5281/zenodo.8347726 ) …”

Association of Histone H3 Trimethylation in Circulating Monocytes with Lack of Early Recurrence in Patients with Bladder Cancer following BCG Induction Therapy

“…To identify peaks that best discriminate between patients who suffered from early disease recurrence and patients with no early recurrence at different time points, we used a machine learning-based feature selection algorithm ( MFeaST ) [ 11 ]. The algorithm relies on a set of machine learning techniques to rank the features based on their ability to discriminate between two groups and is highly effective at identifying molecular patterns among samples.…”