Taking population stratification into account by local permutations in rare-variant association studies on small samples

Mullaert, Jimmy; Bouaziz, Matthieu; Seeleuthner, Yoann; Bigio, Benedetta; Casanova, Jean‐Laurent; Alcaïs, Alexandre; Abel, Laurent; Cobat, Aurélie

doi:10.1101/2020.01.29.924977

Cited by 3 publications

(7 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For large sample sizes, both PC and LMM controlled for stratification better at larger geographic scales than at finer scales. In small samples (50 cases and 100 controls), PC approaches gave inflated type I errors even in the absence of population stratification, as previously reported [18, 29, 39]. This inflation disappeared when the sample included additional controls, whatever their ethnic origin, even with a highly unbalanced case-control ratio.…”

Section: Discussionmentioning

confidence: 53%

“…By contrast, the type I error of LMM was inflated in samples with highly unbalanced case-control ratios, whatever the level of population stratification, as previously noted in the context of GWAS [40]. Finally, the adapted local permutations procedure recently proposed by Mullaert et al [29] gave very promising results, as it fully corrected for population stratification, regardless of the scale over which the stratification occurred, sample size and case-control ratio. When valid under H 0 , the three correction methods had similar powers.…”

Section: Discussionmentioning

confidence: 82%

“…With the classical permutation approach, the simulation of test statistics under H 0 is achieved by randomly resampling phenotypes ( i.e exchanging them between individuals). Adapted local permutations are based on the observation that, in the presence of population structure, not all phenotypes are exchangeable [29]. A given sample has a higher chance of sharing its phenotype with another sample of the same ancestry.…”

Section: Methodsmentioning

confidence: 99%

“…These neighborhoods are established according to a genetic distance derived from the first 10 PCs: where PC CV is the matrix of principal components calculated on the set of common variants and λ k is the eigenvalue corresponding to the k -th principal component . This distance is used to create a neighborhood of 30 individuals around each sample [29]. Permutations can then be performed for each sample, within its neighborhood.…”

Section: Methodsmentioning

confidence: 99%

“…We focused on a classic genetic model for a rare disease with a phenotype driven by rare deleterious variants well suited for a burden test, such as the cohort allelic sums test (CAST) [3]. We tested two classical correction methods, PC and LMM, a promising novel correction method called adapted local permutations (LocPerm) [29] and considered an uncorrected CAST-like test as a reference. Our global objective here is to provide useful practical insight into how best to account for population stratification in rare variant association studies.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Controlling for Human Population Stratification in Rare Variant Association Studies

Bouaziz

Mullaert

Bigio

et al. 2020

Preprint

View full text Add to dashboard Cite

22 Population stratification is a strong confounding factor in human genetic association studies. In 23 analyses of rare variants, the main correction strategies based on principal components (PC) and linear 24 mixed models (LMM), may yield conflicting conclusions, due to both the specific type of structure 25 induced by rare variants and the particular statistical features of association tests. Studies evaluating 26 these approaches generally focused on specific situations with limited types of simulated structure 27 and large sample sizes. We investigated the properties of several correction methods in the context 28 of a large simulation study using real exome data, and several within-and between-continent 29 stratification scenarios. We also considered different sample sizes, with situations including as few 30 as 50 cases, to account for the analysis of rare disorders. In this context, we focused on a genetic 31 model with a phenotype driven by rare deleterious variants well suited for a burden test. For analyses 32 of large samples, we found that accounting for stratification was more difficult with a continental 33 structure than with a worldwide structure. LMM failed to maintain a correct type I error in many 34 scenarios, whereas PCs based on common variants failed only in the presence of extreme continental 35 stratification. When a sample of 50 cases was considered, an inflation of type I errors was observed 36 with PC for small numbers of controls (≤100), and with LMM for large numbers of controls (≥1000).37 We also tested a promising novel adapted local permutation method (LocPerm), which maintained a 38 correct type I error in all situations. All approaches capable of correcting for stratification properly 39 had similar powers for detecting actual associations pointing out that the key issue is to properly 40 control type I errors. Finally, we found that adding a large panel of external controls (e.g. extracted 41 from publicly available databases) was an efficient way to increase the power of analyses including 42 small numbers of cases, provided an appropriate stratification correction was used. 43 44 3 45 Author Summary 4647 Genetic association studies focusing on rare variants using next generation sequencing (NGS) data 48 have become a common strategy to overcome the shortcomings of classical genome-wide association 49 studies for the analysis of rare and common diseases. The issue of population stratification remains 50 however a substantial question that has not been fully resolved when analyzing NGS data. In this 51 work, we propose a comprehensive evaluation of the main strategies to account for stratification, that 52 are principal components and linear mixed model, along with a novel approach based on local 53 permutations (LocPerm). We compared these correction methods in many different settings, 54 considering several types of population structures, sample sizes or types of variants. Our results 55 highlighted important limitations of some classical methods as those using principal com...

show abstract

Section: Discussionmentioning

confidence: 53%

Section: Discussionmentioning

confidence: 82%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Controlling for Human Population Stratification in Rare Variant Association Studies

Bouaziz

Mullaert

Bigio

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

Rare genetic variants explain missing heritability in smoking

et al. 2022

View full text Add to dashboard Cite

Background: Across complex traits, common variants explain only a modest amount of variance, with SNP-heritability consistently below heritability estimates from close relatives. Here, we examined the contribution of rare variant to tobacco use risk in up to 26,000 individuals of European ancestry in the Trans-Omics for Precision Medicine (TOPMed) program with whole genome sequence (WGS;~30X coverage). Method:We grouped about 35million genetic variants by their minor allele frequencies (MAF) and linkage disequilibrium (LD) and estimated SNP-heritability for age of smoking initiation (N=14,747), cigarettes smoked per day (N=15,425), smoking cessation (N=17,871) and initiation (N=26,340) using linear mixed model. Rare variant population structure is detected and adjusted for by permutation procedure. We estimated an upper bound for narrow-sense heritability for tobacco use using available pedigrees consisting of close relatives in TOPMed.Results: Rare variants with MAF 0.1% to 0.01%, mostly from non-protein altering region, accounted for 26% of variation in age of initiation and 15% for cessation. Follow-up analysis indicated that about onethird of these rare variants contribtion is potentially confounded with rare variants structure even after adjusting for principal components. After further conservative adjustment of population structure, we estimated SNP-based heritability to be 0.21 (SE=0.08) for age of initiation, 0.15 (0.06) for cigarettes per day, 0.21 (0.09) for cessation, and 0.24 (0.07) for initiation, 1.8-4.5 times higher than previous SNP-based estimates. Our pedigree-based upper-bound for SNP-based heritability ranged from 0.18-0.35. Conclusion:The substantial contribution of rare variants for several smoking phenotypes sheds light on the missing heritability and genetic etiology of tobacco use. It also informs fine-mapping strategies since the majority of the rare variant contribution was located in non-coding regulatory regions.

show abstract

Genetic architecture of smoking: Evaluating rare variant contribution from deep whole-genome sequencing of up to 26,000 individuals

Jang¹,

Evans²,

Fialkowski³

et al. 2021

Preprint

View full text Add to dashboard Cite

Background Across complex traits, common variants explain only a modest amount of variance, with SNP-heritability consistently below heritability estimates from close relatives. Here, we examined the contribution of rare variant to tobacco use risk in up to 26,000 individuals of European ancestry in the Trans-Omics for Precision Medicine (TOPMed) program with whole genome sequence (WGS;~30X coverage).Method We grouped about 35 million genetic variants by their minor allele frequencies (MAF) and linkage disequilibrium (LD) and estimated SNP-heritability for age of smoking initiation (N = 14,747), cigarettes smoked per day (N = 15,425), smoking cessation (N = 17,871) and initiation (N = 26,340) using linear mixed model. Rare variant population structure is detected and adjusted for by permutation procedure. We estimated an upper bound for narrow-sense heritability for tobacco use using available pedigrees consisting of close relatives in TOPMed.Results Rare variants with MAF 0.1–0.01%, mostly from non-protein altering region, accounted for 26% of variation in age of initiation and 15% for cessation. Follow-up analysis indicated that about one-third of these rare variants contribtion is potentially confounded with rare variants structure even after adjusting for principal components. After further conservative adjustment of population structure, we estimated SNP-based heritability to be 0.21 (SE = 0.08) for age of initiation, 0.15 (0.06) for cigarettes per day, 0.21 (0.09) for cessation, and 0.24 (0.07) for initiation, 1.8–4.5 times higher than previous SNP-based estimates. Our pedigree-based upper-bound for SNP-based heritability ranged from 0.18–0.35.Conclusion The substantial contribution of rare variants for several smoking phenotypes sheds light on the missing heritability and genetic etiology of tobacco use. It also informs fine-mapping strategies since the majority of the rare variant contribution was located in non-coding regulatory regions.

show abstract

Taking population stratification into account by local permutations in rare-variant association studies on small samples

Cited by 3 publications

References 15 publications

Controlling for Human Population Stratification in Rare Variant Association Studies

Controlling for Human Population Stratification in Rare Variant Association Studies

Rare genetic variants explain missing heritability in smoking

Genetic architecture of smoking: Evaluating rare variant contribution from deep whole-genome sequencing of up to 26,000 individuals

Contact Info

Product

Resources

About