Recently, the greatest statistical computational challenge in genetic epidemiology is to identify and characterize the genes that interact with other genes and environment factors that bring the effect on complex multifactorial disease. These gene-gene interactions are also denoted as epitasis in which this phenomenon cannot be solved by traditional statistical method due to the high dimensionality of the data and the occurrence of multiple polymorphism. Hence, there are several machine learning methods to solve such problems by identifying such susceptibility gene which are neural networks (NNs), support vector machine (SVM), and random forests (RFs) in such common and multifactorial disease. This paper gives an overview on machine learning methods, describing the methodology of each machine learning methods and its application in detecting gene-gene and gene-environment interactions. Lastly, this paper discussed each machine learning method and presents the strengths and weaknesses of each machine learning method in detecting gene-gene interactions in complex human disease.
High kernel elongation (HKE) is one of the high-quality characteristics in rice. The objectives of this study were to determine the effects of ageing treatments, gene actions, and inheritance pattern of kernel elongation on cooking quality in two populations of rice and determine the path of influence and contribution of other traits to kernel elongation in rice. Two rice populations derived from crosses between MR219 × Mahsuri Mutan and MR219 × Basmati 370 were used. The breeding materials included two F1 progenies from the two populations, and their respective parents were grown in four different batches at a week interval to synchronize the flowering between the female and male plants. Scaling tests and generation means analysis were carried out to determine ageing effects and estimate additive-dominance gene action and epistasis. The estimation of gene interaction was based on quantitative traits. Path coefficient analysis was done using SAS software version 9.4 to determine the path of influence (direct or indirect) of six quantitative traits on HKE. Results obtained showed that nonallelic gene interaction was observed in all traits. The results before ageing and after ageing showed significant differences in all traits, while the gene interaction changed after ageing. The HKE value improved after ageing, suggesting that ageing is an external factor that could influence gene expression. The epistasis effect for HKE obtained from the cross Mahsuri Mutan × MR219 showed duplicate epistasis while that obtained from a cross between Basmati 370 × MR219 showed complimentary epistasis. Besides, the heritability of HKE was higher in Basmati 370 × MR219 compared to that obtained in Mahsuri Mutan × MR219. The path analysis showed that the cooked grain length and length-width ratio positively significantly affected HKE. It was concluded that ageing treatment is an external factor that could improve the expression of HKE. The findings from this study would be useful to breeders in the selection and development of new specialty (HKE) rice varieties.
The Apolipoprotein E ε4 (APOE ε4) haplotype is the strongest genetic risk factor for late-onset Alzheimer's disease (AD). The Translocase of Outer Mitochondrial Membrane-40 (TOMM40) gene maintains cellular bioenergetics, which is disrupted in AD. TOMM40 rs2075650 ('650) G versus A carriage is consistently related to neural and cognitive outcomes, but it is unclear if and how it interacts with APOE. We examined 21 orthogonal neural networks among 8,222 middle-aged to aged participants in the UK Biobank cohort. ANOVA and multiple linear regression tested main effects and interactions with APOE and TOMM40 '650 genotypes, and if age and sex acted as moderators. APOE ε4 was associated with less strength in multiple networks, while '650 G versus A carriage was related to more language comprehension network strength. In APOE ε4 carriers, '650 G-carriage led to less network strength with increasing age, while in non-G-carriers this was only seen in women but not men. TOMM40 may shift what happens to network activity in aging APOE ε4 carriers depending on sex.
Genetic polymorphism has been implicated as a factor for the occurrence of non-alcoholic fatty liver disease (NAFLD). This study attempted to assess whether polymorphisms in the leptin receptor (LEPR) gene and its combined effect with patatin-like phospholipase domain-containing protein 3 (PNPLA3/adiponutrin) are associated with risk of NAFLD.
Angiotensin II type 1 receptor (AGTR1) has been reported to play a fibrogenic role in non-alcoholic fatty liver disease (NAFLD). In this study, five variants of the AGTR1 gene (rs3772622, rs3772627, rs3772630, rs3772633, and rs2276736) were examined for their association with susceptibility to NAFLD. Subjects made up of 144 biopsy-proven NAFLD patients and 198 controls were genotyped using TaqMan assays. The liver biopsy specimens were histologically graded and scored according to the method of Brunt. Single locus analysis in pooled subjects revealed no association between each of the five variants with susceptibility to NAFLD. In the Indian ethnic group, the rs2276736, rs3772630 and rs3772627 appear to be protective against NAFLD (p = 0.010, p = 0.016 and p = 0.026, respectively). Haplotype ACGCA is shown to be protective against NAFLD for the Indian ethnic subgroup (p = 0.03). Gene-gene interaction between the AGTR1 gene and the patatin-like phospholipase domain-containing 3 (PNPLA3) gene, which we previously reported as associated with NAFLD in this sample, showed a strong interaction between AGTR1 (rs3772627), AGTRI (rs3772630) and PNPLA3 (rs738409) polymorphisms on NAFLD susceptibility (p = 0.007). Further analysis of the NAFLD patients revealed that the G allele of the AGTR1 rs3772622 is associated with increased fibrosis score (p = 0.003). This is the first study that replicates an association between AGTR1 polymorphism and NAFLD, with further details in histological features of NAFLD. There is lack of evidence to suggest an association between any of the five variants of the AGTR1 gene and NAFLD in the Malays and Chinese. In the Indians, the rs2276736, rs3772630 and rs3772627 appear to protect against NAFLD. We report novel findings of an association between the G allele of the rs3772622 with occurrence of fibrosis and of the gene-gene interaction between AGTR1gene and the much-studied PNPLA3 gene.