Biological Feature Selection and Disease Gene Identification using New Stepwise Random Forests

Abstract

Identifying disease genes from human genome is a critical task in biomedical research. Important biological features to distinguish the disease genes from the non-disease genes have been mainly selected based on traditional feature selection approaches. However, the traditional feature selection approaches unnecessarily consider many unimportant biological features. As a result, although some of the existing classification techniques have been applied to disease gene identification, the prediction performance was not satisfactory. A small set of the most important biological features can enhance the accuracy of disease gene identification, as well as provide potentially useful knowledge for biologists or clinicians, who can further investigate the selected biological features as well as the potential disease genes. In this paper, we propose a new stepwise random forests (SRF) approach for biological feature selection and disease gene identification. The SRF approach consists of two stages. In the first stage, only important biological features are iteratively selected in a forward selection manner based on one-dimensional random forest regression, where the updated residual vector is considered as the current response vector. We can then determine a small set of important biological features. In the second stage, random forests classification with regard to the selected biological features is applied to identify disease genes. Our extensive experiments show that the proposed SRF approach outperforms the existing feature selection and classification techniques in terms of biological feature selection and disease gene identification.

Candidate genes associated with neurological manifestations of COVID-19: Meta-analysis using multiple computational approaches

“…The Gene Ontology resources provides a model of hierarchically (ancestors-descendants relationship) organized directed acyclic graph (DAG) having GO-terms as nodes and functional association as directed edges within each hierarchy by ‘ is_a ’ (subtype) and ‘ part_of ’ (component) relationship associated to gene/protein functionality (molecular function, cellular component and biological process) description 73,74 . GO based biological process (GO-BP) provides cohesive evidences about protein interactions related to both physical and functional networks of molecular events involved in cellular physiology 83–85 . The candidate genes for a disease show common biological pathway(s) 48 .…”

Predicting Parkinson’s Disease Related Genes Based on PyFeat and Gradient Boosted Decision Tree (GBDT)

“…As mentioned above, The current studies have several limitations, which can be summarized in the following points. First, most studies have developed methods to predict the genes related to disease, but a few of these methods were designed for PD-genes prediction 13,14,[16][17][18] . Second, some of these PD methods identified only protein genes related to PD and ignored lncRNA genes, whether lncRNAs is critical for improving our understanding and diagnosing different diseases 12,20,22,23 .…”

“…Based on ensemble-weighted classifiers, they built the EPU learning to predict disease-related genes. Hwang 18 proposed stepwise random forests (SRF) method to select the biological features for identifying genes related to the disease. First, they integrated multiple biological features from the gene characteristics, such as protein domains, gene ontology, and human protein interactions.…”

“…Pseudo Composition: This measure determines the frequencies of sub-sequences, where n is the sub-sequences length. The number of generated features from a sequence is defined by equation (18), where n = 3, then only eighty-four (84) features exist. These features are determined by the frequencies of sub-sequences: A, C, G, T, AA, .…”

N-semble-based method for identifying Parkinson’s disease genes

“…Experimental results have confirmed that the classifier built with high-dimensional features may not be an efficient in terms of detection of PD. Hwang [17] proposed a SRF (Stepwise Random Forests) approach for disease gene identification on biological data sources used by Yang. He had enhanced his method by considering only important features with filter-based feature selection method for classification.…”