A robust and stable gene selection algorithm based on graph theory and machine learning

Subrata Saha,Sanguthevar Rajasekaran,Ahmed Soliman

doi:10.1186/s40246-021-00366-9

Subrata Saha, Sanguthevar Rajasekaran + Show 1 more

Open Access

https://doi.org/10.1186/s40246-021-00366-9

Copy DOI

Abstract

BackgroundNowadays we are observing an explosion of gene expression data with phenotypes. It enables us to accurately identify genes responsible for certain medical condition as well as classify them for drug target. Like any other phenotype data in medical domain, gene expression data with phenotypes also suffer from being a very underdetermined system. In a very large set of features but a very small sample size domain (e.g. DNA microarray, RNA-seq data, GWAS data, etc.), it is often reported that several contrasting feature subsets may yield near equally optimal results. This phenomenon is known as instability. Considering these facts, we have developed a robust and stable supervised gene selection algorithm to select a set of robust and stable genes having a better prediction ability from the gene expression datasets with phenotypes. Stability and robustness is ensured by class and instance level perturbations, respectively.ResultsWe have performed rigorous experimental evaluations using 10 real gene expression microarray datasets with phenotypes. They reveal that our algorithm outperforms the state-of-the-art algorithms with respect to stability and classification accuracy. We have also performed biological enrichment analysis based on gene ontology-biological processes (GO-BP) terms, disease ontology (DO) terms, and biological pathways.ConclusionsIt is indisputable from the results of the performance evaluations that our proposed method is indeed an effective and efficient supervised gene selection algorithm.

Highlights

Nowadays we are observing an explosion of gene expression data with phenotypes
Stability and robustness Ensuring stability and robustness mitigates 3 key issues dominating in supervised feature selection domain: (1) In a very underdetermined system where we have few a hundreds to thousands of samples with thousands to millions of features (e.g., DNA microarray, RNA Sequencing (RNA-seq) data, or GWAS data), it is often found that contrasting feature subsets of similar size may yield identical results
In this article, we have proposed a robust and stable supervised gene selection algorithm Robust and Stable Gene Selection Algorithm (RSGSA) based on graph theory and ensembles of linear Support Vector Machine (SVM)

Summary

Introduction

Nowadays we are observing an explosion of gene expression data with phenotypes. In a very large set of features but a very small sample size domain (e.g. DNA microarray, RNA-seq data, GWAS data, etc.), it is often reported that several contrasting feature subsets may yield near optimal results. Traditional statistical methods [5] are designed to analyze susceptibility of genes from gene expression data with phenotype by considering only a single gene at a time. Nowadays with generation sequencing methods (e.g., RNA-seq, CAGE, etc.), specific transcript expression can be identified. The expressions of those transcripts are measured from not more than several thousands of individuals

Methods

Results

Discussion

Conclusion