Multifactor Dimensionality Reduction Method Research Articles

Impact of the gene-gene interactions related to the HIF-1α signaling pathway with the knee osteoarthritis development.

Articular cartilage is the target tissue of osteoarthritis (OA), and because it lacks capillary networks, the microenvironment is hypoxic. Hypoxia inducible factor-1 alpha (HIF-1α) regulates the homeostasis of this tissue. The aim of this study was to investigate whether genetic polymorphisms of the HIF-1α signaling pathway are involved in the development of knee OA. We performed a case-control association study and genotyped 134 knee OA patients and 267 healthy controls. All participants were genotyped in order to evaluate 42 SNPs from 22 genes involved in the HIF-1α signaling pathway using the OpenArray technology. Gene-gene interactions (epistasis) were analyzed using the multifactor dimensionality reduction (MDR) method. The MDR analysis showed epistasis between AKT2 (rs8100018) and IGF1 (rs2288377), AKT2 (rs8100018) and IGF1 (rs35767), IGF1 (rs35767) and COL2A1 (rs1793953), and between GSK3B (rs6438552) and IGF1 (rs35767) polymorphisms, with information gain values of 21.24%, 8.37%, 9.93%, and 5.73%, respectively. Additionally, our model allowed us to identify high- and low-risk genotypes among COL2A1 rs1793953, GSK3B rs6438552, AKT2 rs8100018, and IGF1 rs35767 polymorphisms. Knowing the interactions of these polymorphisms involved in HIF-1α signaling pathway could provide a new diagnostic support tool to identify individuals at high risk of developing knee OA.

Predicting Type 2 Diabetes Incidence with Genome-wide Gene-gene and Gene-diet Interactions (OR31-08-19)

ObjectivesTo identify genetic and dietary factors, and their interactions that contribute to type 2 diabetes (T2D) and predict an individual’s risk to design more precise prevention and treatment strategies. MethodsA genome-wide scan for up to three-way interactions between 717,275 single nucleotide polymorphisms (SNPs), and 139 dietary and lifestyle factors was conducted on 1380 participants of the Boston Puerto Rican Health Study using the Generalized Multifactor Dimensionality Reduction (GMDR) method. Based on identified genetic and dietary factors, we then used machine learning (ML) to predict T2D risk, and the accuracy of prediction was assessed using area under the Receiver Operating Characteristic curve (ROC-AUC). ResultsA genome-wide scan for main effects and up to three-way interactions between SNPs and dietary factors using GMDR identified a set of 818 SNPs and 12 dietary factors that were selected for the prediction of T2D incidence. Comparing several ML algorithms, we found that stochastic gradient boosting provided the best prediction accuracy of T2D incidence with ROC-AUC of 0.93 in the training set, and overall accuracy of 85% based on test set validation. This approach identified that 52 SNPs in 37 genes, three food groups of high sugar content, and age were key predictors of the best-fit model. ConclusionsThis study illustrates a powerful methodology that can predict incidence of T2D based on gene-gene and gene-environment interactions in combination with machine learning. This genome-wide approach allows identification of those diet and lifestyle factors that interact with genotype and can inform personalized nutrition strategies for the prevention and treatment of T2D. Funding SourcesThis work was funded by the US Department of Agriculture, under agreement no. 8050-51000-098-00D, and NIH grants P01 AG023394, P50 HL105185, and R01 AG027087.

Variants in matrix metalloproteinase-9 gene are associated with hemorrhagic transformation in acute ischemic stroke patients with atherothrombosis, small artery disease, and cardioembolic stroke.

ObjectiveThe potential effect of matrix metalloproteinase‐9 (MMP‐9) variants and these variants interactions on hemorrhagic transformation (HT) risk after ischemic stroke (IS) remain unclear. The aims of present study were to investigate the associations of six variants in MMP‐9 with HT, and these variants interactions whether related to increased HT risk.MethodA total of 705 patients with IS who were admitted to the participating hospitals within 48 hr of symptom onset were consecutively enrolled between March 2014 and December 2016. HT was confirmed by brain computed tomography (CT) scan during 14 days from stroke onset. Six variants of MMP‐9 gene were measured by mass spectrometry. Interactions of gene variant–gene variant were assessed through generalized multifactor dimensionality reduction method (GMDR).ResultsHT occurred in 104 (14.8%) patients. There were no differences in genotypes for the six variants between patients with and without HT using univariate analysis (all p > 0.05). GMDR analysis revealed that there was a synergistic effect of gene variant–gene variant interactions between rs3918242 and rs3787268 in MMP‐9 gene. Cox regression analysis showed that high‐risk interactions of rs3918242 and rs3787268 were associated with increased risk of HT after adjusting for covariates (hazard ratio: 2.08; 95% confidence interval: 1.34–7.85; p = 0.016).ConclusionIncidence of HT is common in acute IS in Chinese population. The mechanisms leading to HT are most likely multifactorial. Two‐loci interactions of rs3918242 and rs3787268 in MMP‐9 gene may confer a higher risk for HT.

Interaction Effects of Plasma Vitamins A, E, D, B9, and B12 and Tobacco Exposure in Urothelial Bladder Cancer: A Multifactor Dimensionality Reduction Analysis

The study aimed to examine circulating vitamins A, E, D, and B12 and folate in patients with urothelial bladder cancer (UBC) and detect potential interaction effects of these micronutrients on UBC risk. A case–control study was conducted on 262 UBC patients and 254 matched controls. Vitamins A and E were assessed by ultra performance liquid chromatography, and vitamins D and B12 and folate were assessed by immunological methods. Binary logistic regression models were used to test associations of plasma vitamins tertiles with UBC risk. A multifactor dimensionality reduction method (MDR) was applied to assess interactive effects of the vitamins and tobacco on UBC risk. Higher levels in vitamins A, E, and D were associated with lower occurrence of UBC. No significant association was observed in plasma folate or vitamin B12 with UBC. There were redundancy interactions of plasma vitamin D with tobacco and with plasma vitamin A on UBC risk. Even though the study could not ascertain causality, the findings suggest that vitamins A, E, and D might be protective against UBC. Vitamins A and D interact antagonistically with each other’s and with tobacco to modulate UBC risk. These interactions should be taken in consideration for the prevention of UBC.

Epistatic Interaction Between 5-HT1A and Vascular Endothelial Growth Factor Gene Polymorphisms in the Northern Chinese Han Population With Major Depressive Disorder.

Aims: Serotonin 1A receptor (5-HT1A) and vascular endothelial growth factor (VEGF) are widely expressed in the neurons of the hippocampus and have significant roles in the pathophysiological processes of major depressive disorders (MDDs). The present study was designed to examine 5-HT1A and VEGF gene polymorphisms and whether the gene–gene interaction of 5-HT1A and VEGF gene variants was associated with MDD. Methods: A total of 264 MDD patients and 264 healthy controls were included in the present genetic study. The rs6295, rs1364043, and rs878567 single-nucleotide polymorphisms (SNPs) in the 5-HT1A gene and the rs699947, rs833061, and rs2010963 SNPs in the VEGF gene were selected for genotypic analyses. The generalized multifactor dimensionality reduction method was employed to assess their interactions. Results: The genotype distributions of the two genes’ respective SNPs were significantly different between patients and controls for 5-HT1A rs6295 (p = 0.041) and VEGF rs2010963 (p = 0.035); however, no significant allelic variation in 5-HT1A (rs6295, rs1364043, and rs878567) and VEGF (rs699947, rs833061, and rs2010963) was found. The interactions between 5-HT1A (rs6295, rs1364043, and rs878567) and VEGF (rs699947, rs833061, and rs2010963) had a cross-validation (CV) consistency of 10/10 and a p value of 0.0107, which was considered as the best generalized multifactor dimensionality reduction (GMDR) model. Conclusions: The interactions between 5-HT1A and VEGF gene polymorphisms may play a key role in the development of MDD in the Northern Chinese Han population.

The Interaction of TPH2 and 5-HT2A Polymorphisms on Major Depressive Disorder Susceptibility in a Chinese Han Population: A Case-Control Study.

Purpose: TPH2 and 5-HT2A appear to play vital roles in the homeostatic regulation of serotonin levels in the brain, their genetic variations may lead to impaired homeostatic regulation of serotonin resulting in abnormal levels of serotonin in the brain, thus predisposing individuals to MDD. However, research studies have yet to confirm which gene-gene interaction effect between TPH2 and 5-HT2A polymorphisms results in increased susceptibility to MDD.Methods: A total of 565 participants, consisting of 278 MDD patients and 287 healthy controls from the Chinese Han population, were recruited for the present study. Six single nucleotide polymorphisms (SNPs) of TPH2/5-HT2A were selected to assess their interaction by use of a generalized multifactor dimensionality reduction method.Results: A-allele carriers of rs11178997 and rs120074175 were more likely to suffer from MDD than T-allele carriers of rs11178997, or G-allele carriers of rs120074175. The interaction between TPH2 (rs120074175, rs11178997) and 5-HT2A (rs7997012) was considered as the best multi-locus model upon the MDD susceptibility.Conclusions: Our data identified an important effect of TPH2 genetic variants (rs11178997 and rs120074175) upon the risk of MDD, and suggested that the interaction of TPH2/5-HT2A polymorphism variants confer a greater susceptibility to MDD in Chinese Han population.

Использование метода Multifactor Dimensionality Reduction (MDR) и его модификаций для анализа ген-генных и генно-средовых взаимодействий при генетико-эпидемиологических исследованиях (обзор)

Использование метода Multifactor Dimensionality Reduction (MDR) и его модификаций для анализа ген-генных и генно-средовых взаимодействий при генетико-эпидемиологических исследованиях (обзор)

The TXA2R rs1131882, P2Y1 rs1371097 and GPIIIa rs2317676 three-loci interactions may increase the risk of carotid stenosis in patients with ischemic stroke

BackgroundThe genetic risk factors for carotid stenosis are not fully understood. The aim of this study is to investigate the relationship between variants in platelet activation-relevant genes and carotid stenosis in patients with ischemic stroke (IS).MethodsEleven variants of platelet activation-relevant genes, aggregates of platelet-leukocyte, and platelet aggregation were examined in 236 IS patients with carotid stenosis and 378 patients without carotid stenosis. High-resolution B-mode ultrasound was used to assess carotid stenosis. Generalized multifactor dimensionality reduction (GMDR) methods were applied in analyzing gene-gene interactions to determine whether there was any interactive role of assessed variants in affecting risk of carotid stenosis.ResultsPlatelet aggregation and aggregates of platelet-leukocyte showed higher value in patients with carotid stenosis, compared with patients without carotid stenosis. Excluding potential disturbance variables, these 11 variants were not associated with carotid stenosis. However, according to the GMDR analysis, gene-gene interactions among TXA2R rs1131882, P2Y1 rs1371097 and GPIIIa rs2317676 had a synergistic influence on carotid stenosis. The high-risk interactions between the three variants showed a relationship with higher platelet activation, and have independent associations with risk of carotid stenosis (OR = 2.72, 95% CI: 1.28–7.82, P = 0.001).ConclusionThe interactions among rs1131882, rs1371097 and rs2317676 perhaps increase the risk of symptomatic carotid stenosis, and maybe a potential marker for carotid stenosis. In this study, the combinatorial analysis made good use in elucidating complex risk factors in the heredity of carotid stenosis.

Epistasis between ADIPOQ rs1501299 and PON1 rs662 polymorphisms is potentially associated with the development of knee osteoarthritis.

Overweight produces oxidative stress (OS) on the articular cartilage, with the subsequent risk of developing knee osteoarthritis (OA). Associations between genetic polymorphisms related to OS and OA have been reported, but it is currently unknown whether there exist interactions among them that affect OA development. To identify and evaluate interactions between multiple SNPs related to OS in Mexican knee OA patients. Ninety-two knee OA patients were included in the study, which were compared to 147 healthy controls. Nine variants of six genes (PEPD, AGER, IL6, ADIPOQ, PON1, and CA6) related to OS were genotyped in both study groups through the OpenArray system. Epistasis was analyzed with the multifactor dimensionality reduction (MDR) method. The MDR analysis revealed a significant interaction (p = 0.0107) between polymorphisms rs1501299 (ADIPOQ) and rs662 (PON1), with an entropy value of 9.84%; in addition, high and low risk genotypes were identified between these two polymorphisms. The effect of the interaction between rs1501299 (ADIPOQ) and rs662 (PON1) polymorphisms seems to play an important role in OA pathogenesis; so the epistasis analysis may provide an excellent tool for identifying individuals at high risk for developing OA.

Haplotype Analysis on the Relationship of the DNAJC6 Gene with Early-Onset Parkinson's Disease Risk in a Chinese Population.

DNAJC6 gene is one of the Parkinson's disease (PD) related genes, but relationship between DNAJC6 polymorphisms and PD remains unclear. We aims to examine the association between genetic variations in DNAJC6 gene and idiopathic early-onset PD (EOPD) in the Chinese population. Exons and intron/exon boundaries of DNAJC6 gene was amplified and sequenced in 135 EOPD patients and 212 healthy controls. Single nucleotide polymorphisms (SNP)-based and haplotype-based association study between EOPD and DNAJC6 was conducted. SNP-SNP interactions were investigated using the generalized multifactor dimensionality reduction (GMDR) method. We further evaluated the effect of variants on gene function and expression using online in silico algorithms and databases. We found fourteen previously reported SNPs in the DNAJC6 gene. The frequencies of variant alleles in rs11208644, rs4582839 and rs4915691 were observed significantly higher in EOPD group compared to healthy controls, while in rs6588144 was significantly lower. Additionally, haplotype analysis indicated that the CTCACTCGGC, CTTACTCGGC and TTTGTTCGAC haplotypes were associated with higher EOPD risk in EOPD patients. SNP-SNP interaction analysis showed that rs12077111-rs4592284 SNP combination was the best model with higher EOPD risk. Based on the in silico analysis results, these SNPs were predicted to be no harm to the protein function, but might lead to possible changes in splice site and alter the expression level of DNAJC6. Our study indicated that EOPD was associated with several SNPs and haplotypes of DNAJC6 gene.

Are MMP3, MMP8 and TIMP2 gene variants associated with anterior cruciate ligament rupture susceptibility?

ObjectivesAnterior cruciate ligament rupture (ACLR) is a common and severe knee injury which typically occurs as a result of sports participation, primarily via a non-contact mechanism. A number of extrinsic and intrinsic risk factors, including genetics, have been identified thus far. Matrix metalloproteinases (MMPs) and tissue inhibitors of metalloproteases (TIMPs) play a crucial role in extracellular matrix remodeling of ligaments and therefore the genes encoding MMPs and TIMPs are plausible candidates for investigation with ACL rupture risk. DesignA case-control genetic association study was conducted on 229 (158 male) individuals with surgically diagnosed primary ACLR, ruptured through non-contact mechanisms and 192 (107 male) apparently healthy participants (CON) without any history of ACLR. All participants were physically active, unrelated, self-reported Caucasians. MethodsAll participants were genotyped for four single nucleotide polymorphisms (SNP): MMP3 (rs591058C/T, rs679620 G/A), MMP8 (rs11225395C/T), and TIMP2 (rs4789932 G/A) using standard PCR assays. Gene-gene interactions were inferred. Single-locus association analysis was conducted using the Chi-square test. SNP-SNP interaction effects were analysed using multifactor dimensionality reduction (MDR) method. ResultsGenotype frequencies did not significantly differ between cases and controls, however, the MMP3 rs679620 G and rs591058C alleles were significantly overrepresented in cases compared to controls (p=0.021, OR=1.38, 95% CI: 1.05–1.81). ConclusionsThese results support the hypothesis that genetic variation within MMP3 contributes to inter-individual susceptibility to non-contact ACLR. However, these results need to be explored further in larger, independent sample sets.

SHANK1 polymorphisms and SNP-SNP interactions among SHANK family: A possible cue for recognition to autism spectrum disorder in infant age.

Autism spectrum disorder (ASD) is a serious lifelong neurodevelopmental disorder. ASD is diagnosed for children at the age of two. ASD diagnosis, as early as possible, lays the foundation for treatment and much better prognosis. Notably, gene-based test is an inherent method to recognize the potential infants with ASD before the age of two. To investigate whether SHANK family contributes to ASD prediction, on the basis of our previous studies of SHANK2 and SHANK3, we further investigated associations between SHANK1 polymorphisms and ASD risk as well as SNP-SNP interactions among SHANK family. We enrolled 470 subjects (229 cases and 241 healthy controls) who were northeast Chinese Han. Four tag SNPs (rs73042561, rs3745521, rs4801846, and rs12461427) of SHANK1 were selected and genotyped. We used the SNPStats online analysis program to assess the associations between the four SNPs and ASD risk. The SNP-SNP interactions among SHANK family were analyzed using multifactor dimensionality reduction method. We found that the four SHANK1 SNPs were not associated with ASD risk in northeast Chinese Han population. There existed a strong synergistic interaction between rs11236697 [SHANK2] and rs74336682 [SHANK2], and moderate synergistic interactions (rs74336682 [SHANK2]-rs73042561 [SHANK1], rs11236697 [SHANK2]-rs77716438 [SHANK2], and rs11236697 [SHANK2]-rs75357229 [SHANK2]). These SHANK1 variants may not affect the susceptibility to ASD in Chinese Han population. SNP-SNP interactions in SHANK family may confer ASD risk. Autism Res 2019, 12: 375-383 © 2019 International Society for Autism Research, Wiley Periodicals, Inc. LAY SUMMARY: ASD is a serious lifelong neurodevelopmental disorder with strong genetic components. We investigated associations between SHANK1 polymorphisms and ASD risk as well as SNP-SNP interactions among SHANK family. Our results indicated that there exists no association between SHANK1 SNPs and ASD, and SNP-SNP interactions in SHANK family may confer ASD risk in the Northeast Han Chinese population. Future studies are needed to test more SHANK family SNPs in a large sample to demonstrate the associations.

Unified Cox model based multifactor dimensionality reduction method for gene-gene interaction analysis of the survival phenotype

BackgroundOne strategy for addressing missing heritability in genome-wide association study is gene-gene interaction analysis, which, unlike a single gene approach, involves high-dimensionality. The multifactor dimensionality reduction method (MDR) has been widely applied to reduce multi-levels of genotypes into high or low risk groups. The Cox-MDR method has been proposed to detect gene-gene interactions associated with the survival phenotype by using the martingale residuals from a Cox model. However, this method requires a cross-validation procedure to find the best SNP pair among all possible pairs and the permutation procedure should be followed for the significance of gene-gene interactions. Recently, the unified model based multifactor dimensionality reduction method (UM-MDR) has been proposed to unify the significance testing with the MDR algorithm within the regression model framework, in which neither cross-validation nor permutation testing are needed. In this paper, we proposed a simple approach, called Cox UM-MDR, which combines Cox-MDR with the key procedure of UM-MDR to identify gene-gene interactions associated with the survival phenotype.ResultsThe simulation study was performed to compare Cox UM-MDR with Cox-MDR with and without the marginal effects of SNPs. We found that Cox UM-MDR has similar power to Cox-MDR without marginal effects, whereas it outperforms Cox-MDR with marginal effects and more robust to heavy censoring. We also applied Cox UM-MDR to a dataset of leukemia patients and detected gene-gene interactions with regard to the survival time.ConclusionCox UM-MDR is easily implemented by combining Cox-MDR with UM-MDR to detect the significant gene-gene interactions associated with the survival time without cross-validation and permutation testing. The simulation results are shown to demonstrate the utility of the proposed method, which achieves at least the same power as Cox-MDR in most scenarios, and outperforms Cox-MDR when some SNPs having only marginal effects might mask the detection of the causal epistasis.

Generalized multifactor dimensionality reduction approaches to identification of genetic interactions underlying ordinal traits.

The manifestation of complex traits is influenced by gene-gene and gene-environment interactions, and the identification of multifactor interactions is an important but challenging undertaking for genetic studies. Many complex phenotypes such as disease severity are measured on an ordinal scale with more than two categories. A proportional odds model can improve statistical power for these outcomes, when compared to a logit model either collapsing the categories into two mutually exclusive groups or limiting the analysis to pairs of categories. In this study, we propose a proportional odds model-based generalized multifactor dimensionality reduction (GMDR) method for detection of interactions underlying polytomous ordinal phenotypes. Computer simulations demonstrated that this new GMDR method has a higher power and more accurate predictive ability than the GMDR methods based on a logit model and a multinomial logit model. We applied this new method to the genetic analysis of low-density lipoprotein (LDL) cholesterol, a causal risk factor for coronary artery disease, in the Multi-Ethnic Study of Atherosclerosis, and identified a significant joint action of the CELSR2, SERPINA12, HPGD, and APOB genes. This finding provides new information to advance the limited knowledge about genetic regulation and gene interactions in metabolic pathways of LDL cholesterol. In conclusion, the proportional odds model-based GMDR is a useful tool that can boost statistical power and prediction accuracy in studying multifactor interactions underlying ordinal traits.

High genetic risk scores of SLIT3, PLEKHA5 and PPP2R2C variants increased insulin resistance and interacted with coffee and caffeine consumption in middle-aged adults

Backgrounds and AimsInsulin resistance is a common feature of metabolic syndrome that may be influenced by genetic risk factors. We hypothesized that genetic risk scores (GRS) of SNPs that influence insulin resistance and signaling interact with lifestyles to modulate insulin resistance in Korean adults. Methods and resultsGenome-wide association studies (GWAS) of subjects aged 40–65 years who participated in the Ansung/Ansan cohorts (8842 adults) in Korea revealed 52 genetic variants that influence insulin resistance. The best gene-gene interaction model was explored using the generalized multifactor dimensionality reduction (GMDR) method. GRS from the best model were calculated and the GRS were divided into low, medium and high groups. The best model for representing insulin resistance included SLIT3_rs2974430, PLEKHA5_rs1077044, and PPP2R2C_rs16838853. The odds ratios for insulin resistance were increased by 150% in the High-GRS group compared to the Low-GRS group. However, ORs for insulin secretion capacity, measured by HOMA-B, were not associated with GRS. Coffee and caffeine intake and GRS had an interaction with insulin resistance: In subjects with high coffee (≥10 cups/week) or caffeine intake (≥220 mg caffeine/day), insulin resistance was significantly elevated in the High-GRS group, but not in the Low-GRS. However, alcohol intake, smoking and physical activity did not have an interaction with GRS. Insulin secretion capacity was not significantly influenced by GRS when evaluating the adjusted odds ratios. ConclusionsSubjects with High-GRS may be susceptible to increased insulin resistance by 50% and its risk may be exacerbated by consuming more than 10 cups coffee/week or 220 mg caffeine/day.

Dvl3 polymorphism interacts with life events and pro-inflammatory cytokines to influence major depressive disorder susceptibility

The purpose of this study is to explore Dvl3 variants and their interaction with negative life events on MDD susceptibility in a Chinese Han population. Additionally, we also attempted to identify whether there is an association between Dvl3 variants and pro-inflammatory cytokines. A total of 1102 participants, consisting of 550 patients with MDD and 552 healthy subjects, were recruited for genotyping by TaqMan allelic discrimination assay. Pro-inflammatory cytokine mRNA levels in peripheral blood were measured by QPCR. After the assessment of negative life events by the Life Events Scale, the Dvl3 gene–environment interaction (G × E) and risk factors were evaluated using generalized multifactor dimensionality reduction method (GMDR) and logistic regression analysis, respectively. This study is the first to reveal the interaction between Dvl3 allelic variations and negative life events as well as pro-inflammatory cytokines on MDD susceptibility in a Chinese Han population.

Associations of Homer Scaffolding Protein 1 gene and psychological correlates with suicide attempts in Chinese: A pilot study of multifactorial risk model

BackgroundSuicide is a fatal outcome for subjects with mental ill-health. Genetic factors and psychological correlates are believed to contribute to the risk of suicide attempts (SA), whereas both factors are reported to exert a small effect. This study therefore tried to investigate if combination of the two aspects can enhance the explanation of variance in SA. MethodsA common variant rs7713917 in HOMER1 gene was genotyped for 333 Chinese psychiatric patients with or without SA. Multifactorial risk models comprised of this variant and psychological correlates were identified by logistic regression analysis (LRA) and Multifactor Dimensionality Reduction (MDR) method separately, and then evaluated for their performance by biostatistical methods. ResultsAn association of A-carrier genotypes in rs7713917 with an increased risk of SA was observed (OR = 1.79, 95% CI = 1.08–2.98). Although with a medium effect size, this variant alone could only explain 1.9% variance of SA. Interestingly, this study was the first time to show that the association of the rs7713917 and SA was significantly mediated by the NEO conscientiousness (NEOC) dimension (p-value = 0.002), with a greater genetic effect observed in subjects with a low NEO-C level (OR = 2.89, 95% CI = 1.16–7.18) but not in subjects with an average or high level. Upon the LRA method, the multifactorial risk model constituted of the two interacted factors and their interaction effect could explain up to 17.0% variance, which was almost 9-fold higher than the one explained by the rs7713917 alone. Furthermore, this model owned a higher effectiveness than the three models identified by the MDR method (p-value < 0.0007). ConclusionsThe present study identified an effective multifactorial risk model, in which the combination of the HOMER1 variant and the NEO-C dimension could enhance explanation of the variance of SA in Chinese. This pilot study may provide a novel avenue to investigate the pathogenesis of SA in psychiatric patients.

Class Balanced Multifactor Dimensionality Reduction to Detect Gene–Gene Interactions

Detecting gene-gene interactions in single-nucleotide polymorphism data is vital for understanding disease susceptibility. However, existing approaches may be limited by the sample size in case-control studies. Herein, we propose a balance approach for the multifactor dimensionality reduction (BMDR) method to increase the accuracy of estimates of the prediction error rate in small samples. BMDR explicitly selects the best model by evaluating the average of prediction error rates over k-fold cross-validation without cross-validation consistency selection. In this study, we used several epistatic models with and without marginal effects under different parameter settings (heritability and minor allele frequencies) to evaluate the performance of existing approaches. Using simulated data sets, BMDR successfully detected gene-gene interactions, particularly for data sets with small sample sizes. A large data set was obtained from the Wellcome Trust Case Control Consortium, and results indicated that BMDR could effectively detect significant gene-gene interactions.

Genetic variants of MGMT, RHPN2, and FAM49A contributed to susceptibility of nonsyndromic orofacial clefts in a Chinese population.

The role of underlying genetic factors in the pathogenesis of nonsyndromic orofacial clefts (NSOC) remains poorly understood. Although genomewide association studies (GWASs) of NSOC have successfully identified a large number of novel genetic risk loci, association results of replication studies are inconsistent across different populations. Six single nucleotide polymorphisms (SNPs) (rs7922405 at 10q26.3, rs73039426 at 19q13.11, rs7552 at 2p24.2, rs1788160 at 8q22.2, rs9381107 at 6p24.3, and rs17095681 at 10q25.3) were analyzed for an association with NSOC in 1062 participants of Chinese descent (596 patients and 466 controls). We applied the multifactor dimensionality reduction (MDR) method to detect potential gene-gene (G × G) interactions in the six SNPs. The genotype or allele frequencies of SNPs rs7922405, rs73039426, and rs7552 showed significant differences between the controls and patients with NSOC, whereas no association was shown between three SNPs (rs1788160, rs17095681, and rs9381107) and NSOC. MDR analysis did not reveal significant G × G interactions for susceptibility to NSOC. We confirmed that three genes (rs7922405 of MGMT, rs73039426 of RHPN2, and rs7552 of FAM49A) may contribute to NSOC in Chinese populations. MGMT and RHPN2 are associated with NSOC, which is herein demonstrated for the first time.

Abstract 2350: Efficient survival multifactor dimensionality reduction method for genome-wide association study

Abstract The problem of identifying associations between SNPs and case-control outcome has been studied extensively and a number of new techniques have been developed. Little progress has been made, however in the analysis of SNP effect in relation to censored survival data. We present an extension of the two class multifactor dimensionality reduction (MDR) algorithm that enables detection and characterization of epistatic SNP-SNP interactions in the context of survival outcome. The proposed an Efficient Survival MDR (ES-MDR) method handles censored data by modifying MDR's constructive induction algorithm to use logrank Test. We applied ES-MDR to genetic data of over 470,000 SNPs from the OncoArray Consortium. We use onset age of lung cancer and case-control (n=27,312) status as the survival outcome and divided data into training and testing sets. We also adjust for subject's age, gender and smoking status. From training set, we identified strong association from SNPs in BRCA1 and IL17RC genes with lung cancer onset age. This result is validated in the testing set. Citation Format: Jiang Gui, Xuemei Ji, Christopher I. Amos. Efficient survival multifactor dimensionality reduction method for genome-wide association study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2350.