A tutorial on conducting genome‐wide association studies: Quality control and statistical analysis

Marees, Andries T.; Kluiver, Hilde de; Stringer, Sven; Vorspan, Florence; Curis, Emmanuel; Marie-Claire, Cynthia; Derks, Eske M.

doi:10.1002/mpr.1608

Cited by 569 publications

(500 citation statements)

References 51 publications

(58 reference statements)

Supporting

Mentioning

458

Contrasting

Unclassified

Order By: Relevance

“…However, it is overly strict for densely genotyped and imputed studies where correlations between variants exist (Power et al, 2016b) and requires much larger sample sizes in order to detect causal variants. To overcome the issue of strictness, some tools (Jaillard et al, 2018) implement the Benjamini Hochberg false discovery rate (FDR) (Benjamini and Hochberg, 1995), a less stringent method to control for multiple testing Type I errors but it has also been found to be conservative (Storey and Tibshirani, 2003) as it assumes that SNPs are independent, which is seldom true (Marees et al, 2018). Understanding the level of LD between SNPs and computing an appropriate significance threshold that is optimal for each study (Visscher et al, 2012) therefore presents a feasible and ideal solution.…”

Section: Analytical Considerations and Pitfallsmentioning

confidence: 99%

Current Affairs of Microbial Genome-Wide Association Studies: Approaches, Bottlenecks and Analytical Pitfalls

et al. 2020

View full text Add to dashboard Cite

Microbial genome-wide association studies (mGWAS) are a new and exciting research field that is adapting human GWAS methods to understand how variations in microbial genomes affect host or pathogen phenotypes, such as drug resistance, virulence, host specificity and prognosis. Several computational tools and methods have been developed or adapted from human GWAS to facilitate the discovery of novel mutations and structural variations that are associated with the phenotypes of interest. However, no comprehensive, end-to-end, user-friendly tool is currently available. The development of a broadly applicable pipeline presents a real opportunity among computational biologists. Here, (i) we review the prominent and promising tools, (ii) discuss analytical pitfalls and bottlenecks in mGWAS, (iii) provide insights into the selection of appropriate tools, (iv) highlight the gaps that still need to be filled and how users and developers can work together to overcome these bottlenecks. Use of mGWAS research can inform drug repositioning decisions as well as accelerate the discovery and development of more effective vaccines and antimicrobials for pressing infectious diseases of global health significance, such as HIV, TB, influenza, and malaria.

show abstract

Section: Analytical Considerations and Pitfallsmentioning

confidence: 99%

Current Affairs of Microbial Genome-Wide Association Studies: Approaches, Bottlenecks and Analytical Pitfalls

et al. 2020

View full text Add to dashboard Cite

show abstract

“…We did not impute any genotypes to prevent false positive associations and a larger multiple testing burden. There were 551,839 typed SNPs; subsequent SNP and individual filtering and trimming was based on 1) SNPs with > 20% missing data (239 removed), 2) individuals with > 20% missing data (0 removed), 3) minor allele frequency < 0.01 to remove rare variant associations (260,269 removed), 4) SNPs out of Hardy Weinberg equilibrium for quantitative traits (58 removed due to P<1e -6 ) (40). All samples passed kinship and heterozygosity thresholds after the filtering outlined above, leaving 62 samples and 291,273 SNPs to analyse.…”

Section: Methodsmentioning

confidence: 99%

Identification of novel mitochondrial and mitochondrial related genetic loci associated with exercise response in the Gene SMART study

Harvey

Voisin

Lea

et al. 2020

Preprint

View full text Add to dashboard Cite

20Mitochondria supply intracellular energy requirements during exercise. Specific mitochondrial haplogroups and 21 mitochondrial genetic variants have been associated with athletic performance, and exercise responses. However, 22 these associations were discovered using underpowered, candidate gene approaches, and consequently have not 23 been replicated. Here, we used whole-mitochondrial genome sequencing, in conjunction with high-throughput 24 genotyping arrays, to discover novel genetic variants associated with exercise responses in the Gene SMART 25(Skeletal Muscle Adaptive Response to Training) cohort (n=62 completed). We performed a Principal Component 26 Analysis of cohort aerobic fitness measures to build composite traits and test for variants associated with exercise 27 outcomes. None of the mitochondrial genetic variants but nine nuclear encoded variants in eight separate genes 28 were found to be associated with exercise responses (FDR<0.05) (rs11061368: DIABLO, rs113400963: 29 FAM185A, rs6062129 and rs6121949: MTG2, rs7231304: AFG3L2, rs2041840: NDUFAF7, rs7085433: 30 TIMM23, rs1063271: SPTLC2, rs2275273: ALDH18A1). Additionally, we outline potential mechanisms by 31 which these variants may be contributing to exercise phenotypes. Our data suggest novel nuclear-encoded SNPs 32 and mitochondrial pathways associated with exercise response phenotypes. Future studies should focus on 33 validating these variants across different cohorts and ethnicities. 34 35 AUTHOR SUMMARY 36Previous exercise genetic studies contain many flaws that impede the growth in knowledge surrounding change 37 in exercise outcomes. In particular, exercise studies looking at mtDNA variants have looked at very small portions 38 of the mitochondrial genome. Mitochondria are the 'power house' of the cell and therefore understanding the 39 mitochondrial genetics behind adaptations to training can help us fill knowledge gaps in current research. Here, 40we utilised a new mitochondrial genetic sequencing technique to examine all mitochondrial and mitochondrial 41 related genetic variations. We have shown that there were no mitochondrial specific variants that influenced 42 exercise training however there were 9 related variants that were significantly associated with exercise 43 phenotypes. Additionally, we have shown that building composite traits increased the significance of our 44 association testing and lead to novel findings. We will be able to understand why response to training is so varied 45 and increase the effectiveness of exercise training on a host of metabolic disorders. 48Responses to exercise training depends on the type of exercise stimulus, and varies considerably between 49 individuals (1-3). This variability is tissue-specific, and may be explained by a combination of genetic variants, 50 epigenetic signatures, other molecular and lifestyle factors (4, 5). Mitochondria are the key mediators of 51 intracellular energy and are involved in many essential cell metabolism and homeostasis processes (6) with 52 exercise trai...

show abstract

“…Quality Control (QC) was done in two steps: 1) the missingness threshold for SNPs was set to 0.2 and 28,610 SNPs were excluded; 2) two thresholds for the Hardy-Weinberg equilibrium were used: 10 −10 was the P-value threshold for excluding SNPs in HP cases, and 10 −6 was the P-value threshold for excluding SNPs in controls 14 . Ten SNPs in HP cases and 10 SNPs in controls failed the test.…”

Section: Methodsmentioning

confidence: 99%

Using Genome-Wide Association Study to Identify Genes and Pathways associated with Hypersensitivity Pneumonitis

Millerick

Rosenman

Cui

et al. 2020

Preprint

View full text Add to dashboard Cite

BackgroundHypersensitivity Pneumonitis (HP) is an interstitial lung disease caused by an immune response to the inhalation of antigens. Since only a small proportion of individuals exposed to HP-related antigens develop the disease, a genetic variation may play a role in disease development.MethodsIn this small-scale study, 24 patients diagnosed with HP were matched with control group who shared the patient’s environment and were exposed to the same HP-associated antigens. Logistic regression was employed to identify Single-Nucleotide Polymorphisms (SNPs) associated with HP. Next genes associated with HP were identified using sequence kernel association test (SKAT) analysis. Last, Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Oncology (GO) enrichment analysis were employed to find HP signaling pathways using SNPs coded on genes and on non-coding genes, respectively.ResultsGiven the small sample size, no single SNPs or genes were identified to be significantly associated with HP after adjustment for multiple testing. After P-value adjustment, the KEGG and GO pathway enrichment analysis identified 11 and 20 significant pathways respectively using SNPs coded on genes. Among these pathways, Cell cycle, Proteasome and Base excision repair had previously reported to be associated with lung function.ConclusionThis is the first GWAS study identifying genetic factors associated with HP. Although no significant associations at SNPs/gene level were identified, there were significant pathways that are identified associated with HP which need further investigation in large cohorts.

show abstract

A tutorial on conducting genome‐wide association studies: Quality control and statistical analysis

Cited by 569 publications

References 51 publications

Current Affairs of Microbial Genome-Wide Association Studies: Approaches, Bottlenecks and Analytical Pitfalls

Current Affairs of Microbial Genome-Wide Association Studies: Approaches, Bottlenecks and Analytical Pitfalls

Identification of novel mitochondrial and mitochondrial related genetic loci associated with exercise response in the Gene SMART study

Using Genome-Wide Association Study to Identify Genes and Pathways associated with Hypersensitivity Pneumonitis

Contact Info

Product

Resources

About