Search bioRxiv⌕ Search

Biology subjects

Gourraud, P.-A.

Publications and source records attributed to Gourraud, P.-A..

7 recordsLinked to original sources

Optimal HLA imputation of admixed population with dimension reduction

Human genomics has quickly evolved, powering genome-wide association studies (GWASs). SNP-based GWASs cannot capture the intense polymorphism of HLA genes, highly associated with disease susceptibility. There are methods to statistically impute HLA genotypes from SNP-genotypes data, but lack of diversity in reference panels hinders their performance. We evaluated the accuracy of the 1,000 Genomes data as a reference panel for imputing HLA from admixed individuals of African and European ancestries, focusing on (a) the full dataset, (b) 10 replications from 6 populations, (c) 19 conditions for the custom reference panels. The full dataset outperformed smaller models, with a good F1-score of 0.66 for HLA-B. However, custom models outperformed the multiethnic or population models of similar size (F1-scores up to 0.53, against up to 0.42). We demonstrated the importance of using genetically specific models for imputing admixed populations, which are currently underrepresented in public datasets, opening the door to HLA imputation for every genetic population.

genomics↗

DrBERT: A Robust Pre-trained Model in French for Biomedical and Clinical domains

In recent years, pre-trained language models (PLMs) achieve the best performance on a wide range of natural language processing (NLP) tasks. While the first models were trained on general domain data, specialized ones have emerged to more effectively treat specific domains. In this paper, we propose an original study of PLMs in the medical domain on French language. We compare, for the first time, the performance of PLMs trained on both public data from the web and private data from healthcare establishments. We also evaluate different learning strategies on a set of biomedical tasks. In particular, we show that we can take advantage of already existing biomedical PLMs in a foreign language by further pre-train it on our targeted data. Finally, we release the first specialized PLMs for the biomedical field in French, called DrBERT, as well as the largest corpus of medical data under free license on which these models are trained.

bioinformatics↗

The Nature of Genetic Susceptibility to Multiple Sclerosis

OBJECTIVETo explore the nature of MS-susceptibility and, by extension, other complex-genetic diseases. BACKGROUNDBasic-epidemiological parameters of MS (e.g., prevalence, recurrence-risks for siblings and twins, time-dependent changes in sex-ratio, etc.) are well-established. Moreover, >200 genetic-loci are unequivocally MS-associated, especially the HLA-DRB1*15:01~HLA-DQB1*06:02~a1 haplotype-association. DESIGN/METHODSWe define the "genetically-susceptible" subset-(G) to include everyone with any non-zero life-time chance of developing MS. We analyze, mathematically, the implications that these epidemiological observations have regarding genetic susceptibility. In addition, we use the sex-ratio change (observed over a 35-year interval), to derive the relationship between MS-probability and an increasing likelihood of a suitable environmental-exposure. RESULTSWe demonstrate that genetic-susceptibitly is restricted to less than 4.7% of populations across Europe and North America. Among carriers of the HLA-DRB1*15:01~HLA-DQB1*06:02~a1 haplotype, fewer than 20% are even in the subset-(G). Women are less likely to be susceptible than men although their MS-penetrance is considerably greater. Response-curves for MS-probability increase with an increasing likelihood of a suitable environmental-exposure, especially among women. These environmental response-curves plateau at under 50% for women and at a significantly lower level for men. CONCLUSIONSMS is fundamentally a genetic disorder. Despite this, a suitable environmental-exposure is also critical for disease-pathogenesis. Genetic-susceptibility requires specific combinations of non-additive genetic risk-factors. For example, the HLA-DRB1*15:01~HLA-DQB1*06:02~a1 haplotype, by itself, poses no MS-risk. Moreover, the fact that environmental-response-curves plateau below 50%, indicates that disease-pathogenesis is partly stochastic. By extension, other diseases for which monozygotic-twin recurrence-risks greatly exceed disease-prevalence (e.g., rheumatoid arthritis, diabetes, and celiac disease), must have a similar genetic basis. Author SummaryWe define a "genetic-susceptible" subset (G) of the general population (Z) to include everyone with any non-zero chance of developing MS over their life-time. Using well-established epidemiological data from across Europe and North America, we establish that genetic-susceptibility is confined to less than 4.7% of these populations. Thus, the large majority of individuals have no chance whatsoever of developing MS, irrespective of any environmental conditions that they may experience during their lifetimes. In this sense, MS is fundamentally a genetic disorder. And, indeed, more than 200 genetic-loci, in multiple genomic locations have now been well-established to be associated with MS. Notably, however, the HLA-DRB1*15:01~HLA-DQB1*06:02~a1 or (H+) haplotype, which has, by far, the strongest MS-association of any, has a carrier frequency in the population of 23% in North America and Europe. Therefore, with genetic susceptibility in the population being less than 4.7%, more than 80% of (H+)-haplotype carriers, must not be genetically-susceptible and, thus, have no chance of developing MS. In this circumstance, genetic susceptibility to MS must arise from a combination of this haplotype with "susceptible states" at other genetic loci. By itself, the (H+)-haplotype poses no risk. Indeed, genetic-susceptibility, generally, seems to require specific combinations of non-additive genetic risk-factors. Naturally, the conclusion that MS is fundamentally genetic does not preclude the possibility the environmental events are also critical to disease-pathogenesis. Using epidemiological data about the world-wide increase in the (F:M) sex-ratio for MS to construct (for men and women separately) the response curves relating an increasing likelihood of MS to an increasing likelihood of a sufficient environmental exposure (i.e., an exposure sufficient to cause MS in a susceptible individual). This analysis provides insight to both disease-susceptibility and disease-pathogenesis. First, men are more likely to be susceptible than women although susceptible women are considerably more likely to actually develop MS. Second, men seem to have a lower environmental threshold than women for developing MS. Nevertheless, women are more responsive to changes in the environmental conditions compared to men. Third, even with a maximal environmental exposure, susceptible women never exceed a 50% chance of developing MS. By contrast, susceptible men have a significantly lower likelihood (<10% chance) of developing MS. This indicates that stochastic factors must also be critical in disease pathogenesis. Finally, the nature of genetic susceptibility developed here for MS is applicable to many other complex genetic disorders. Indeed, for any disease, in which the proband-wise MZ-twin concordance rate greatly exceeds the disease-prevalence in the population (e.g., type I diabetes, rheumatoid arthritis, and celiac disease), only a small fraction of the population can possibly be genetically susceptible, as defined.

genetics↗

Genetic Susceptibility to Multiple Sclerosis: Interactions between Conserved Extended Haplotypes of the MHC and other Susceptibility Regions

OBJECTIVETo study the accumulation of MS-risk resulting from different combinations of MS-associated conserved-extended-haplotypes of the MHC and three non-MHC risk-loci nearby genes EOMES, ZFP36L1, CLEC16A.\n\nBACKGROUNDDefining \"genetic-susceptibility\" as having a non-zero probability of developing MS, both theoretical considerations and epidemiological observations indicate that only 2.2-4.5% of northern-populations can possibly be \"genetically-susceptible\" to MS. Nevertheless, many haplotypes (both within the MHC and elsewhere) are unequivocally MS-associated and, yet, have population-frequencies of >20%. Such frequency-disparities underscore the complex-interactions that must occur between these \"risk-haplotypes\" and MS-susceptibility.\n\nDESIGN/MEHTODSThe WTCCC dataset was statistically-phased at the MHC and at three other susceptibility-regions. Haplotypes were stratified by their impact on \"MS-risk\". MS-associations for different combinations of \"risk-haplotypes\" were assessed. The appropriateness of both additive and multiplicative risk-accumulation models was determined.\n\nRESULTSCombinations of different \"risk-haplotypes\" produced an MS-risk that was considerably closer to an additive model than a multiplicative model. Nevertheless, neither of these simple probability-models adequately accounted for the accumulation of disease-risk in MS at these four loci.\n\nCONCLUSIONS\"Genetic-susceptibility\" to MS seems to depend upon the exact state at each \"risk-locus\" and upon specific gene-gene combinations across loci. Moreover, \"genetic-susceptibility\" is both rare in the population and, yet, is a necessary condition for MS to develop in any individual. In this sense, MS is a \"genetic\" disease. Nevertheless although, \"genetic-susceptibility\" is a necessary condition for MS to develop, environmental factors (whatever these may be) and stochastic processes are also necessary determinants of whether a \"genetically-susceptible\" individual will actually get MS.\n\nAuthor SummaryDefining a \"genetically-susceptible\" individual to be any person in the population who has any chance of developing multiple sclerosis (MS), we demonstrate that, at a theoretical level and using widely-accepted epidemiological observations, only 2.2-4.5% of individuals in northern populations can possibly be \"genetically susceptible\" to MS. Thus, more than 95.5% of individuals in these populations have no chance of getting MS, regardless of the environmental circumstances that they may experience.\n\nNevertheless, certain \"susceptibility-haplotypes\" (e.g., HLA-DRB1*15:01~DQB1*06:02) have a far greater carrier-frequency than 2.2-4.5%. Consequently, most carriers of these \"susceptibility-haplotypes\" have no chance of getting MS and, therefore, their \"susceptibility\" must arise from some combination of these haplotypes with other \"susceptibility-haplotypes\". By analyzing such combinatorial impacts at four susceptibility-loci, we found significant interactions both within and between the different \"susceptibility-haplotypes\", thereby confirming the relationship between \"genetic-susceptibility\" and specific gene-gene combinations.\n\nThe nature of \"genetic-susceptibility\" developed here is applicable to other complex genetic disorders. Indeed, any disease for which the MZ-twin concordance rate is substantially greater than the life-time risk in the general population, only a small fraction of the population can possibly be in the \"genetically-susceptible\" subset (i.e., have any chance of developing the disease).

epidemiology↗

Low frequency and rare coding variation contributes to multiple sclerosis risk

Multiple sclerosis is a common, complex neurological disease, where almost 20% of risk heritability can be attributed to common genetic variants, including >230 identified by genome-wide association studies (Patsopoulos et al., 2017). Multiple strands of evidence suggest that the majority of the remaining heritability is also due to the additive effects of individual variants, rather than epistatic interactions between these variants, or mutations exclusive to individual families. Here, we show in 68,379 cases and controls that as much as 5% of this heritability is explained by low-frequency variation in gene coding sequence. We identify four novel genes driving MS risk independently of common variant signals, which highlight a key role for regulatory T cell homeostasis and regulation, IFN{gamma} biology and NF{kappa}B signaling in MS pathogenesis. As low-frequency variants do not show substantial linkage disequilibrium with other variants, and as coding variants are more interpretable and experimentally tractable than non-coding variation, our discoveries constitute a rich resource for dissecting the pathobiology of MS.

genetics↗

Genome-wide association study of asthma in individuals of African ancestry reveals novel asthma susceptibility loci

BACKGROUNDAsthma is a complex disease with striking disparities across racial and ethnic groups, which may be partly attributable to genetic factors. One of the main goals of the Consortium on Asthma among African-ancestry Populations in the Americas (CAAPA) is to discover genes conferring risk to asthma in populations of African descent.\n\nMETHODSWe performed a genome-wide meta-analysis of asthma across 11 CAAPA datasets (4,827 asthma cases and 5,397 controls), genotyped on the African Diaspora Power Chip (ADPC) and including existing GWAS array data. The genotype data were imputed up to a whole genome sequence reference panel from n=880 African ancestry individuals for a total of 61,904,576 SNPs. Statistical models appropriate to each study design were used to test for association, and results were combined using the weighted Z-score method. We also used admixture mapping as a complementary approach to identify loci involved in asthma pathogenesis in subjects of African ancestry.\n\nRESULTSSNPs rs787160 and rs17834780 on chromosome 2q22.3 were significantly associated with asthma (p=6.57 x 10-9 and 2.97 x 10-8, respectively). These SNPs lie in the intergenic region between the Rho GTPase Activating Protein 15 (ARHGAP15) and Glycosyltransferase Like Domain Containing 1 (GTDC1) genes. Four low frequency variants on chromosome 1q21.3, which may be involved in the \"atopic march\" and which are not polymorphic in Europeans, also showed evidence for association with asthma (1.18 x10-6 [&le;] p [&le;] 3.06 x10-6). SNP rs11264909 on chromosome 1q23.1, close to a region previously identified by the EVE asthma meta-analysis as having a putative African ancestry specific effect, only showed differences in counts in subjects homozygous for alleles of African ancestry. Admixture mapping also identified a significantly associated region on chromosome 6q23.2, which includes the Transcription Factor 21 (TCF21) gene, previously shown to be differentially expressed in bronchial tissues of asthmatics and non-asthmatics.\n\nCONCLUSIONSWe have identified a number of novel asthma association signals warranting further investigation.

bioinformatics↗

Identifying tagging SNPs for African specific genetic variation from the African Diaspora Genome

A primary goal of The Consortium on Asthma among African-ancestry Populations in the Americas (CAAPA) is to develop an African Diaspora Power Chip (ADPC), a genotyping array consisting of tagging SNPs, useful in comprehensively identifying African specific genetic variation. This array is designed based on the novel variation identified in 642 CAAPA samples of African ancestry with high coverage whole genome sequence data (~30x depth). This novel variation extends the pattern of variation catalogued in the 1000 Genomes and Exome Sequencing Projects to a spectrum of populations representing the wide range of West African genomic diversity. These individuals from CAAPA also comprise a large swath of the African Diaspora population and incorporate historical genetic diversity covering nearly the entire Atlantic coast of the Americas. Here we show the results of designing and producing such a microchip array. This novel array covers African specific variation far better than other commercially available arrays, and will enable better GWAS analyses for researchers with individuals of African descent in their study populations. A recent study1 cataloging variation in continental African populations suggests this type of African-specific genotyping array is both necessary and valuable for facilitating large-scale GWAS in populations of African ancestry.

genomics↗