Exploration of errors in variance caused by using the first-order approximation in Mendelian randomization

Mendelian randomization (MR) uses genetic variation as a natural experiment to investigate the causal effects of modifiable risk factors (exposures) on outcomes. Two-sample Mendelian randomization (2SMR) is widely used to measure causal effects between exposures and outcomes via genome-wide association studies. 2SMR can increase statistical power by utilizing summary statistics from large consortia such as the UK Biobank. However, the first-order term approximation of standard error is commonly used when applying 2SMR. This approximation can underestimate the variance of causal effects in MR, which can lead to an increased false-positive rate. An alternative is to use the second-order approximation of the standard error, which can considerably correct for the deviation of the first-order approximation. In this study, we simulated MR to show the degree to which the first-order approximation underestimates the variance. We show that depending on the specific situation, the first-order approximation can underestimate the variance almost by half when compared to the true variance, whereas the second-order approximation is robust and accurate.

The impact of school closures on adolescent health-related outcomes during the COVID-19 pandemic: A natural experiment in South Korea

A large number of countries implemented school closure as a response to the COVID-19 pandemic in 2020. As existing studies mostly rely on retrospective or pre/post comparisons that are prone to unmeasured confounding, the effect of school closure on adolescent health is poorly understood. The South Korean government implemented school closure to prevent the spread of COVID-19. A difference-in-differences comparing changes in health-related outcomes between provinces with differing degrees of school closure was performed. The main analysis group consists of middle school students of age 14 to 16 who were hit hardest where up to 73% of total schooling was taken online in Seoul (the physical attendance was reduced from 170 days to 45 days). For sensitivity analysis, a placebo group of high school students of age 19 who attended all school-days physically was included to detect any violation of our identification strategy. In the main analysis group of boys that experienced reduced physical school-days, both total and vigorous physical activity were reduced (-0.35 [-0.54 -0.17] days/week for vigorous physical activity and -0.38 [-0.61 -0.16] days/week for total physical activity) while such effect was absent in the placebo group of boys that actually did not experience school closure (-0.08 [-0.49 0.32] days/week for vigorous PA and -0.16 [-0.67 0.34] days/week for total PA). In girls, vigorous physical activity decreased (-0.22 [-0.40 -0.04] days/week) but the total physical activity was nearly constant (0.03 [-0.18 0.25] days/week). Other outcomes were largely unchanged.

A genome wide association study for lung function in the Korean population using an exome array

Background/Aims@#Lung function is an objective indicator of diagnosis and prognosis of respiratory diseases. Many common genetic variants have been associated with lung function in multiple ethnic populations. We looked for coding variants associated with forced expiratory volume in 1 second (FEV1) and FEV1/forced vital capacity (FVC) in the Korean general population. @*Methods@#We carried out exome array analysis and lung function measurements of the FEV1 and FEV1/FVC in 7,524 individuals of the Korean population. We evaluated single variants with minor allele frequency greater than 0.5%. We performed look-ups for candidate coding variants associations in the UK Biobank, SpiroMeta, and CHARGE consortia. @*Results@#We identified coding variants in the SMIM29 (C6orf1) (p = 1.2 × 10–5) and HMGA1 locus on chromosome 6p21, the GIT2 (p = 6.5 × 10–5) locus on chromosome 12q24, and the ARHGEF40 (p = 9.9 × 10–5) locus on chromosome 14q11 as having a significant association with lung function (FEV1). We also confirmed a previously reported association with lung function and chronic obstructive pulmonary disease in the FAM13A (p = 4.54 × 10–6) locus on chromosome 4q22, in TNXB (p = 1.30 × 10–6) and in AGER (p = 1.09 × 10–8) locus on chromosome 6p21. @*Conclusions@#Our exome array analysis identified that several protein coding variants were associated with lung function in the Korean population. Common coding variants in SMIM29 (C6orf1), HMGA1, GIT2, FAM13A, TNXB, AGER and low-frequency variant in ARHGEF40 potentially affect lung function, which warrant further study.

Analysis of differences in human leukocyte antigen between the two Wellcome Trust Case Control Consortium control datasets

The Wellcome Trust Case Control Consortium (WTCCC) study was a large genome-wide association study that aimed to identify common variants associated with seven diseases. That study combined two control datasets (58C and UK Blood Services) as shared controls. Prior to using the combined controls, the WTCCC performed analyses to show that the genomic content of the control datasets was not significantly different. Recently, the analysis of human leukocyte antigen (HLA) genes has become prevalent due to the development of HLA imputation technology. In this project, we extended the between-control homogeneity analysis of the WTCCC to HLA. We imputed HLA information in the WTCCC control dataset and showed that the HLA content was not significantly different between the two control datasets, suggesting that the combined controls can be used as controls for HLA fine-mapping analysis based on HLA imputation.

MergeReference: A Tool for Merging Reference Panels for HLA Imputation

Recently developed computational methods allow the imputation of human leukocyte antigen (HLA) genes using intergenic single nucleotide polymorphism markers. To improve the imputation accuracy in HLA imputation, it is essential to increase the sample size and the diversity of alleles in the reference panel. Our software, MergeReference, helps achieve this goal by providing a streamlined pipeline for combining multiple reference panels into one.

Comparison of Two Meta-Analysis Methods: Inverse-Variance-Weighted Average and Weighted Sum of Z-Scores

The meta-analysis has become a widely used tool for many applications in bioinformatics, including genome-wide association studies. A commonly used approach for meta-analysis is the fixed effects model approach, for which there are two popular methods: the inverse variance-weighted average method and weighted sum of z-scores method. Although previous studies have shown that the two methods perform similarly, their characteristics and their relationship have not been thoroughly investigated. In this paper, we investigate the optimal characteristics of the two methods and show the connection between the two methods. We demonstrate that the each method is optimized for a unique goal, which gives us insight into the optimal weights for the weighted sum of z-scores method. We examine the connection between the two methods both analytically and empirically and show that their resulting statistics become equivalent under certain assumptions. Finally, we apply both methods to the Wellcome Trust Case Control Consortium data and demonstrate that the two methods can give distinct results in certain study designs.

Association of HLA Genotype and Fulminant Type 1 Diabetes in Koreans

Fulminant type 1 diabetes (T1DM) is a distinct subtype of T1DM that is characterized by rapid onset hyperglycemia, ketoacidosis, absolute insulin deficiency, and near normal levels of glycated hemoglobin at initial presentation. Although it has been reported that class II human leukocyte antigen (HLA) genotype is associated with fulminant T1DM, the genetic predisposition is not fully understood. In this study we investigated the HLA genotype and haplotype in 11 Korean cases of fulminant T1DM using imputation of whole exome sequencing data and compared its frequencies with 413 participants of the Korean Reference Panel. The HLA-DRB1*04:05-HLA-DQB1*04:01 haplotype was significantly associated with increased risk of fulminant T1DM in Fisher's exact test (odds ratio [OR], 4.11; 95% confidence interval [CI], 1.56 to 10.86; p = 0.009). A histidine residue at HLA-DRbeta1 position 13 was marginally associated with increased risk of fulminant T1DM (OR, 2.45; 95% CI ,1.01 to 5.94; p = 0.054). Although we had limited statistical power, we provide evidence that HLA haplotype and amino acid change can be a genetic risk factor of fulminant T1DM in Koreans. Further large-scale research is required to confirm these findings.