Biology subjects

Aschard, H.

Publications and source records attributed to Aschard, H..

At least 19 recordsLinked to original sources

bioRxiv2026

Evolution of the genetic architecture of uterine disorders

Uterine disorders and menstrual abnormalities are prevalent reproductive conditions with significant clinical consequences. Recent genome-wide association studies have identified hundreds of variants contributing to different uterine disorders, while epidemiological evidence suggests that these disorders co-occur more frequently than expected, implying shared genetic mechanisms. However, the specific genetic variants and biological pathways underlying this shared architecture remain poorly characterized. Here, we conducted a uterus-centric, multi-trait genome-wide association analysis across ten uterine disorders in a European cohort to elucidate this shared genetic architecture. Further, we embedded this architecture in a functional and population genomics framework to investigate its plausible biological mechanisms and its evolutionary history in humans. We confirm strong positive genetic correlations between major uterine disorders and identify 31 independent susceptibility loci jointly affecting the genetic risk of multiple uterine pathologies, substantiating an intertwined biological basis. Populational analyses demonstrated that several of these susceptibility variants exhibit pronounced allele frequency differentiation across global populations suggestive of recent polygenic selection, including variants with well-supported regulatory functions at the ESR1-CCDC170, WNT4, SFR1, FOXO1, ITPR1, DMRT1 and CDKN2B loci. Notably, we show that most derived alleles acquired during recent human evolution increase risk across multiple uterine disorders and may evolve under antagonistic selection. These findings provide functional annotation and population-level prioritization of genetic variants influencing multiple uterine disorders. They also highlight how past evolutionary histories may contribute to population differences in uterine disease prevalence and pathogenesis.

genetics↗

bioRxiv2026

Taxonomic-free metagenome GWAS to identify gut microbiome functions influencing host phenotypes

Genome-wide association studies (GWAS) have been pivotal for uncovering the genetics of human phenotypes, and there is now growing interest in applying GWAS-like methods to explore the role of the microbiome in human health. Here, we developed a taxonomy free GWAS approach that uses k-mers, i.e. DNA words of length k that can capture single nucleotide polymorphisms, insertion or deletion events, and gene presence-absence, to interrogate the gut microbiome. We applied this method to 26 traits spanning demographics, physiological measurements, health, and lifestyle in 938 healthy participants from the Milieu Interieur cohort. We generated a k-mer abundance matrix encompassing 97 million distinct k-mers. GWAS of the 26 traits identified significant associations for seven of them: age, sex, depression, appetite, cooked meat consumption, soda intake, and smoking. By modeling the correlation structure among k-mers, we identified a modest number of independent signals and conducted a comprehensive in silico functional annotation of these signals, revealing potential mechanisms of host-microbiota interaction. Overall, our analyses demonstrate that k-mers can capture biologically relevant functions shared across multiple taxa and provide a refined modeling framework that complements the standard taxonomic-based screening approach.

microbiology↗

bioRxiv2025

Identification and Functional Characterization of Regulatory Variants in DPP9 Associated with COVID-19 Severity

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection leads to a wide-range of clinical outcomes, which have been extensively studied through genome-wide association studies (GWAS). Starting from lead genetic variants associated with COVID-19 infection and severity, we identified a subset of non-coding candidate variants with potential regulatory functions. Using bioinformatics analysis and functional screening in three cell lines, we prioritized two DPP9 variants within a haplotype that increases the risk of severe COVID-19. This haplotype exhibited increased regulatory activity and altered transcription factor binding, suggesting its role in influencing COVID-19 severity through modulation of DPP9 expression in immune and lung cell types. The interest of our study lies in the functional characterization of regulatory variants responsible for the higher levels of DPP9 and lung damage observed in patients with severe COVID-19. These findings advance our understanding of genetic risk factors for COVID-19 and highlight functional SNPs that may guide future therapeutic research.

genetics↗

bioRxiv2023

Trait selection strategy in multi-trait GWAS: Boosting SNPs discoverability

Since the first Genome-Wide Association Studies (GWAS), thousands of variant-trait associations have been discovered. However, the sample size required to detect additional variants using standard univariate association screening is increasingly prohibitive. Multi-trait GWAS offers a relevant alternative: it can improve statistical power and lead to new insights about gene function and the joint genetic architecture of human phenotypes. Although many methodological hurdles of multi-trait testing have been discussed, the strategy to select trait, among overwhelming possibilities, has been overlooked. In this study, we conducted extensive multi-trait tests using JASS (Joint Analysis of Summary Statistics) and assessed which genetic features of the analysed sets were associated with anincreased detection of variants as compared to univariate screening. Our analyses identified multiple factors associated with the gain in the association detection in multi-trait tests. Together, these factors of the analysed sets are predictive of the gain of the multi-trait test (Pearsons {rho} equal to 0.43 between the observed and predicted gain, P < 1.6 x 10-60). Applying an alternative multi-trait approach (MTAG, multi-trait analysis of GWAS), we found that in most scenarios but particularly those with larger numbers of traits, JASS outperformed MTAG. Finally, we benchmark several strategies to select set of traits including the prevalent strategy of selecting clinically similar traits, which systematically underperformed selecting clinically heterogenous traits or selecting sets that issued from our data-driven models. This work provides a unique picture of the determinant of multi-trait GWAS statistical power and outline practical strategies for multi-trait testing.

genetics↗

bioRxiv2023

A multivariate outcome test of covariance

Multivariate analysis is becoming central in studies investigating high-throughput molecular data, yet, some important features of these data are seldom explored. Here, we present MANOCCA (Multivariate Analysis of Conditional CovAriance), a powerful method to test for the effect of a predictor on the covariance matrix of a multivariate outcome. The proposed test is by construction orthogonal to tests based on the mean and variance, and is able to capture effects that are missed by both approaches. We first compare the performances of MANOCCA with existing correlation-based methods and show that MANOCCA is the only test correctly calibrated in simulation mimicking omics data. We then investigate the impact of reducing the dimensionality of the data using principal component analysis when the sample size is smaller than the number of pairwise covariance terms analysed. We show that, in many realistic scenarios, the maximum power can be achieved with a limited number of components. Finally, we apply MANOCCA to 1,000 healthy individuals from the Milieu Interieur cohort, to assess the effect of health, lifestyle and genetic factors on the covariance of two sets of phenotypes, blood biomarkers and flow cytometry-based immune phenotypes. Our analyses identify significant associations between multiple factors and the covariance of both omics data.

bioinformatics↗

bioRxiv2023

Polygenic risk score prediction accuracy convergence

Polygenic risk scores (PRS) trained from genome-wide association study (GWAS) results are set to play a pivotal role in biomedical research addressing multifactorial human diseases. The prospect of using these risk scores in clinical care and public health is generating both enthusiasm and controversy, with varying opinions about strengths and limitations across experts1. The performances of existing polygenic scores are still limited, and although it is expected to improve with increasing sample size of GWAS and the development of new powerful methods, it remains unclear how much prediction can be ultimately achieved. Here, we conducted a retrospective analysis to assess the progress in PRS prediction accuracy since the publication of the first large-scale GWASs using six common human diseases with sufficient GWAS data. We show that while PRS accuracy has grown rapidly for years, the improvement pace from recent GWAS has decreased substantially, suggesting that further increasing GWAS sample size may translate into very modest risk discrimination improvement. We next investigated the factors influencing the maximum achievable prediction using recently released whole genome-sequencing data from 125K UK Biobank participants, and state-of-the-art modeling of polygenic outcomes. Our analyses point toward increasing the variant coverage of PRS, using either more imputed variants or sequencing data, as a key component for future improvement in prediction accuracy.

genetics↗

bioRxiv2023

Multi-trait GWAS for diverse ancestries: Mapping the knowledge gap

BackgroundApproximately 95% of samples analyzed in univariate genome-wide association studies (GWAS) are of European ancestry. This bias toward European ancestry populations in association screening also exists for other analyses and methods that are often developed and tested on European ancestry only. However, existing data in non-European populations, which are often of modest sample size, could benefit from innovative approaches as recently illustrated in the context of polygenic risk scores. MethodsHere, we extend and assess the potential limitations and gain of our multi-trait GWAS pipeline, JASS (Joint Analysis of Summary Statistics), for the analysis of non-European ancestries. To this end, we conducted the joint GWAS of 19 hematological traits and glycemic traits across five ancestries (European (EUR), admixed American (AMR), African (AFR), East Asian (EAS), South-East Asian (SAS)). ResultsWe detected 367 new genome-wide significant associations in non-European populations (15 in Admixed American (AMR), 72 in African (AFR) and 280 in East Asian (EAS)). New associations detected represent 5%, 17% and 13% of associations in the AFR, AMR and EAS populations, respectively. Overall, multi-trait testing increases the replication of European associated loci in non-European ancestry by 15%. Pleiotropic effects were highly similar at significant loci across ancestries (e.g. the mean correlation between multi-trait genetic effects of EUR and EAS ancestries was 0.88). For hematological traits, strong discrepancies in multitrait genetic effects are tied to known evolutionary divergences: the ARKC1 loci, which is adaptive to overcome the p.vivax induced malaria. ConclusionsMulti-trait GWAS can be a valuable tool to narrow the genetic knowledge gap between European and non-European populations.

genetics↗

bioRxiv2023

Quantile Regression for biomarkers in the UK Biobank

Genome-wide association studies (GWAS) for biomarkers important for clinical phenotypes can lead to clinically relevant discoveries. GWAS for quantitative traits are based on simplified regression models modeling the conditional mean of a phenotype as a linear function of genotype. An alternative and easy to apply approach is quantile regression that naturally extends linear regression to the analysis of the entire conditional distribution of a phenotype of interest by modeling conditional quantiles within a regression framework. Quantile regression can be applied efficiently at biobank scale using standard statistical packages in much the same way as linear regression, while having some unique advantages such as identifying variants with heterogeneous effects across different quantiles, including non-additive effects and variants involved in gene-environment interactions; accommodating a wide range of phenotype distributions with invariance to trait transformation; and overall providing more detailed information about the underlying genotype-phenotype associations. Here, we demonstrate the value of quantile regression in the context of GWAS by applying it to 39 quantitative traits in the UK Biobank (n > 300, 000 individuals). Across these 39 traits we identify 7,297 significant loci, including 259 loci only detected by quantile regression. We show that quantile regression can help uncover replicable but unmodelled gene-environment interactions, and can provide additional key insights into poorly understood genotype-phenotype correlations for clinically relevant biomarkers at minimal additional cost.

genomics↗

bioRxiv2023

ChoruMM: a versatile multi-components mixed model for bacterial-GWAS

Genome-wide Association Studies (GWAS) have been central to studying the genetics of complex human outcomes, and there is now tremendous interest in implementing GWAS-like approaches to study pathogenic bacteria. A variety of methods have been proposed to address the complex linkage structure of bacterial genomes, however, some questions remain about to optimize the genetic modelling of bacteria to decipher causal variations from correlated ones. Here we examined the genetic structure underlying whole-genome sequencing data from 3,824 Listeria monocytogenes strains, and demonstrate that the standard human genetics model, commonly assumed by existing bacterial GWAS methods, is inadequate for studying such highly structured organisms. We leverage these results to develop ChoruMM, a robust and powerful approach that consists of a multi-component linear mixed model, where components are inferred from a hierarchical clustering of the bacteria genetic relatedness matrix. Our ChoruMM approach also includes post-processing and visualization tools that address the pervasive long-range correlation observed in bacteria genome and allow to assess the type I error rate calibration.

genetics↗

bioRxiv2022

A high-throughput sequencing approach identifies immunotherapeutic targets for bacterial meningitis in neonates

BackgroundWorldwide, Escherichia coli is the leading cause of neonatal Gram-negative bacterial meningitis, but full understanding of the pathogenesis of this disease is not yet achieved. Moreover, to date, no vaccine is available against bacterial neonatal meningitis. MethodsHere, we used Transposon Sequencing of saturated banks of mutants (TnSeq) to evaluate E. coli K1 genetic fitness in murine neonatal meningitis. We identified E. coli K1 genes encoding for factors important for systemic dissemination and brain infection, and focused on products with a likely outer-membrane or extra-cellular localization, as these are potential vaccine candidates. We used in vitro and in vivo models to study the efficacy of active and passive immunization. ResultsWe selected for further study the conserved surface polysaccharide Poly-{beta}-(1-6)-N-Acetyl Glucosamine (PNAG), as a strong candidate for vaccine development. We found that PNAG was a virulence factor in our animal model. We showed that both passive and active immunization successfully prevented and/or treated meningitis caused by E. coli K1 in neonatal mice. We found an excellent opsonophagocytic killing activity of the antibodies to PNAG and in vitro these antibodies were also able to decrease binding, invasion and crossing of E. coli K1 through two blood brain barrier cell lines. Finally, to reinforce the potential of PNAG as a vaccine candidate in bacterial neonatal meningitis, we demonstrated that Group B Streptococcus, the main cause of neonatal meningitis in developed countries, also produced PNAG and that antibodies to PNAG could protect in vitro and in vivo against this major neonatal pathogen. InterpretationAltogether, these results indicate the utility of a high-throughput DNA sequencing method to identify potential immunotherapy targets for a pathogen, including in this study a potential broad-spectrum target for prevention of neonatal bacterial infections. FundingsANR Seq-N-Vaq, Charles Hood Foundation, Hearst Foundation. Groupe Pasteur Mutualite

microbiology↗

bioRxiv2022

Enhancer/gene relationships: need for more reliable genome-wide reference sets

Differences in cells functions arise from differential action of regulatory elements, in particular enhancers. Like promoters, enhancers are genomic regions bound by transcription factors (TF) that activate the expression of one or several genes by getting physically close to them in the 3D space of the nucleus. As there is increasing evidence that variants associated with common diseases are located in enhancers active in cell types relevant to these diseases, knowing the set of enhancers and more importantly the sets of genes activated by each enhancer (the so-called enhancer/gene or E/G relationships) in a cell type, will certainly help understanding these diseases. There are three broad approaches for the genome-wide identification of E/G relationships in a cell type: (1) genetic link methods or eQTL, (2) functional link methods based on 1D functional data such as open chromatin, histone mark and gene expression and (3) spatial link methods based on 3D data such as HiC. Since (1) and (3) are costly, there has been a focus on developing functional link methods and using data from (1) and (3) to evaluate them, however there is still no consensus on the best functional link method to date. For this reason we decided to start from the two latest benchmarks of the field, namely from the CRISPRi-FlowFISH (CRiFF) technique and from 3D and eQTL data in BENGI, and to evaluate the two methods claimed to be the best one on each of these benchmark studies, namely the ABC model and the Average-Rank method respectively, on the other methods reference data. Not only did we manage to reproduce the results of the two benchmarks but we also saw that none of the two methods performed best on the two reference data. While CRiFF reference data are very reliable, it is not genome-wide and is mostly available on a cancer cell type. On the other hand BENGI is genome-wide but may contain many false positives. This study therefore calls for new reliable and genome-wide E/G reference data rather than new functional link E/G identification methods.

bioinformatics↗

bioRxiv2020

Fitting penalized regressions on very large genetic data using snpnet and bigstatsr

Both R packages snpnet and bigstatsr allow for fitting penalized regressions on individual-level genetic data as large as the UK Biobank. Here we benchmark bigstatsr against snpnet for fitting penalized regressions on large genetic data. We find bigstatsr to be an order of magnitude faster than snpnet when applied to the UK Biobank data (from 4.5x to 35x). We also discuss the similarities and differences between the two packages, provide theoretical insights, and make recommendations on how to fit penalized regressions in the context of genetic data.

genetics↗

bioRxiv2020

Multitrait genetic-phenotype associations to connect disease variants and biological mechanisms

BackgroundGenome-wide association studies (GWAS) uncovered a wealth of associations between common variants and human phenotypes. These results, widely shared across the scientific community as summary statistics, fostered a flurry of secondary analysis: heritability and genetic correlation assessment, pleiotropy characterization and multitrait association test. Amongst these secondary analyses, a rising new field is the decomposition of multitrait genetic effects into distinct profiles of pleiotropy. ResultsWe conducted an integrative analysis of GWAS summary statistics from 36 phenotypes to decipher multitrait genetic architecture and its link to biological mechanisms. We started by benchmarking multitrait association tests on a large panel of phenotype sets and established the Omnibus test as the most powerful in practice. We detected 322 new associations that were not previously reported by univariate screening. Using independent significant associations, we investigated the breakdown of genetic association into clusters of variants harboring similar multitrait association profile. Focusing on two subsets of immunity and metabolism phenotypes, we then demonstrate how SNPs within clusters can be mapped to biological pathways and disease mechanisms, providing a putative insight for numerous SNPs with unknown biological function. Finally, for the metabolism set, we investigate the link between gene cluster assignment and success of drug targets in random control trials. We report additional uninvestigated drug targets classified by clusters. ConclusionsMultitrait genetic signals can be decomposed into distinct pleiotropy profiles that reveal consistent with pathways databases and random control trials. We propose this method for the mapping of unannotated SNPs to putative pathways.

genetics↗

bioRxiv2020

Multi-ancestry genome-wide gene-sleep interactions identify novel loci for blood pressure

Long and short sleep duration are associated with elevated blood pressure (BP), possibly through effects on molecular pathways that influence neuroendocrine and vascular systems. To gain new insights into the genetic basis of sleep-related BP variation, we performed genome-wide gene by short or long sleep duration interaction analyses on four BP traits (systolic BP, diastolic BP, mean arterial pressure, and pulse pressure) across five ancestry groups using 1 degree of freedom (1df) interaction and 2df joint tests. Primary multi-ancestry analyses in 62,969 individuals in stage 1 identified 3 novel loci that were replicated in an additional 59,296 individuals in stage 2, including rs7955964 (FIGNL2/ANKRD33) showing significant 1df interactions with long sleep duration and rs73493041 (SNORA26/C9orf170) and rs10406644 (KCTD15/LSM14A) showing significant 1df interactions with short sleep duration (Pint < 5x10-8). Secondary ancestry-specific two-stage analyses and combined stage 1 and 2 analyses additionally identified 23 novel loci that need external replication, including 3 and 5 loci showing significant 1df interactions with long and short sleep duration, respectively (Pint < 5x10-8). Multiple genes mapped to our 26 novel loci have known functions in sleep-wake regulation, nervous and cardiometabolic systems. We also identified new gene by long sleep interactions near five known BP loci ([≤]1Mb) including NME7, FAM208A, MKLN1, CEP164, and RGL3/ELAVL3 (Pint < 5x10-8). This study indicates that sleep and primary mechanisms regulating BP may interact to elevate BP level, suggesting novel insights into sleep-related BP regulation.

genomics↗

bioRxiv2019

MGMM: an R package for fitting Gaussian Mixture Models on Incomplete Genomics Data

Although missing data are prevalent in applications, existing implementations of Gaussian mixture models (GMMs) require complete data. Standard practice is to perform complete case analysis or imputation prior to model fitting. Both approaches have serious drawbacks, potentially resulting in biased and unstable parameter estimates. Here we present MGMM, an R package for fitting GMMs in the presence of missing data. Using three case studies on real and simulated data sets, we demonstrate that, when the underlying distribution is near-to a GMM, MGMM is more effective at recovering the true cluster assignments than state of the art imputation followed by standard GMM. Moreover, MGMM provides an accurate assessment of cluster assignment uncertainty even when the generative distribution is not a GMM. This assessment may be used to identify unassignable observations. MGMM is available as an R package on CRAN: https://CRAN.R-project.org/package=MGMM.

bioinformatics↗

bioRxiv2019

Estimating the effective sample size in association studies of quantitative traits

The effective sample size (ESS) is a quantity estimated in genome-wide association studies (GWAS) with related individuals and/or linear mixed models used in analysis. ESS originally measured relative power in family-based GWAS and has recently become important for correcting GWAS summary statistics in post-GWAS analyses. However, existing ESS approaches have been overlooked and based on empirical estimation. This work presents an analytical form of ESS in mixed-model GWAS of quantitative traits, which is derived using the expectation of quadratic form and validated in extensive simulations. We illustrate the performance and relevance of our ESS estimator in common GWAS scenarios and analytically show that (i) family-based studies are consistently underpowered compared to studies of unrelated individuals of the same sample size; (ii) conditioning on polygenic genetic effect by linear mixed models boosts power; and (iii) power of detecting gene-environment interaction can be substantially gained or lost in family-based designs depending on exposure distribution. We further analyze UK Biobank dataset in two samples of 336,347 unrelated and 68,910 related individuals. Analysis in unrelated individuals reveals a high accuracy of our ESS estimator compared to the existing empirical approach; and analysis of related individuals suggests that the loss in effective sample size due to relatedness is at most 0.94x. Overall, we provide an analytical form of ESS for guiding GWAS designs and processing summary statistics in post-GWAS analyses.

genetics↗

bioRxiv2019

JASS: Command Line and Web interface for the joint analysis of GWAS results

Genome Wide Association Study (GWAS) has been the driving force for identifying association between genetic variants and human phenotypes. Thousands of GWAS summary statistics covering a broad range of human traits and diseases are now publicly available, and studies have demonstrated their utility for a range of secondary analyses. This includes in particular the joint analysis of multiple GWAS to identify new genetic variants missed by univariate screenings. However, although several methods have been proposed, there are very few large scale applications published so far because of challenges in implementing these methods on real data. Here, we present JASS (Joint Analysis of Summary Statistics), a polyvalent Python package that addresses this need. Our package solves all practical and computational barriers for large-scale multivariate analysis of GWAS summary statistics. This includes data cleaning and harmonization tools, an efficient algorithm for fast derivation of various joint statistics, an optimized data management process, and a web interface for exploration purposes. Benchmark analyses confirmed the strong performances of JASS. We also performed multiple real data analyses demonstrating the strong potential of JASS for the detection of new associated genetic variants across various scenarios. Our package is freely available at https://gitlab.pasteur.fr/statistical-genetics/jass.

bioinformatics↗

bioRxiv2019

Deriving stratified effects from joint models investigating Gene-Environment Interactions

BackgroundModels including an interaction term and performing a joint test of SNP and/or interaction effect are often used to discover Gene-Environment (GxE) interactions. When the environmental exposure is a binary variable, analyses from exposure-stratified models which consist of estimating genetic effect in unexposed and exposed individuals separately can be of interest. In large-scale consortia focusing on GxE interactions in which only the joint test has been performed, it may be challenging to get summary statistics from both exposure-stratified and marginal (i.e not accounting for interaction) models.\n\nResultsIn this work, we developed a simple framework to estimate summary statistics in each stratum of a binary exposure and in the marginal model using summary statistics from the \"joint\" model. We performed simulation studies to assess our estimators accuracy and examined potential sources of bias, such as correlation between genotype and exposure and differing phenotypic variances within exposure strata. Results from these simulations highlight the high theoretical accuracy of our estimators and yield insights into the impact of potential sources of bias. We then applied our methods to real data and demonstrate our estimators retained accuracy after filtering SNPs by sample size to mitigate potential bias.\n\nConclusionsThese analyses demonstrated the accuracy of our method in estimating both stratified and marginal summary statistics from a joint model of gene-environment interaction. In addition to facilitating the interpretation of GxE screenings, this work could be used to guide further functional analyses. We provide a user-friendly Python script to apply this strategy to real datasets. The Python script and documentation are available at https://gitlab.pasteur.fr/statistical-genetics/J2S.

bioinformatics↗