epiBAT: Multi-objective bat algorithm for detection of epistatic interactions

Sitarčík, Jozef; Lucká, Mária

doi:10.1109/informatics47936.2019.9119310

Cited by 2 publications

(1 citation statement)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An Epistatic Interaction Multi-Objective Artificial Bee Colony Algorithm Based on Decomposition (EIMOABC/D) model was suggested for epistasis interaction detection [13]. The multi-objective bat optimization algorithm called epiBat is also presented for epistasis identification [14]. In 2019, Liyan Sun et al [9] developed a multi-revolutionary method called SEE to recognize epistasis effects.…”

Section: Introductionmentioning

confidence: 99%

KMeans-NM-SalpEpi: Genetic Interactions Detection through K-Means Clustering with Nelder-Mead and Salp Optimization Techniques in Genome-Wide Association Studies

Priya

Manavalan

2021

Artificial Intelligence Evolution

View full text Add to dashboard Cite

Complex diseases identification through Gene-Gene Interactions (GGIs) plays a significant challenge in Genome-Wide Association Studies (GWAS). A typical indicator of genetic variations in many human diseases is Single Nucleotide Polymorphisms (SNPs). SNPs are the most prevalent sort of genetic variation seen in human beings. The interactions between various SNPs are called Epistasis or genetic interactions. This research paper proposes a two-stage epistasis detection approach based on K-Means clustering and optimization techniques to detect epistasis effects responsible for complex human diseases. In the screening stage, K-Means clustering is adapted to partition the genotype dataset into various clusters. Traditional K-Means clustering algorithms have the flaw of arbitrary selection of the initial k centroid, which leads to inconsistent solutions and traps in the local optimum. We present a hybridized technique based on the K-Means algorithm and Nelder-Mead (NM) optimization (KMeans-NM) to avoid local optima, and all the genotype data falls into a unique collection of clusters for different runs. In the search stage, Salp Optimization with single objective functions (Salp-SO) and Salp Optimization with multi-objective functions (Salp-MO) are employed over the clusters obtained from the screening stage to find disease correlated SNP combinations. The performance of the various proposed algorithms is tested over the simulated datasets. Experimental findings indicated that the KMeans-NM-SalpEpi-SO and KMeans-NM-SalpEpi-MO method is superior to other techniques.

show abstract

Section: Introductionmentioning

confidence: 99%