bioRxiv Science⌕ Search
Home/Archive/bioRxiv

bioRxiv · 10.1101/002931

LD Score Regression Distinguishes Confounding from Polygenicity in Genome-Wide Association Studies

Brendan Bulik-Sullivan·Po-Ru Loh·Hilary Finucane·Stephan Ripke·Jian Yang·Schizophrenia Working Group Psychiatric Genomics Consortium·Nick Patterson·Mark J Daly·Alkes L Price·Benjamin M Neale
View original source ↗Export metadata ↓

Abstract

Both polygenicity1,2 (i.e. many small genetic effects) and confounding biases, such as cryptic relatedness and population stratification3, can yield inflated distributions of test statistics in genome-wide association studies (GWAS). However, current methods cannot distinguish between inflation from bias and true signal from polygenicity. We have developed an approach that quantifies the contributions of each by examining the relationship between test statistics and linkage disequilibrium (LD). We term this approach LD Score regression. LD Score regression provides an upper bound on the contribution of confounding bias to the observed inflation in test statistics and can be used to estimate a more powerful correction factor than genomic control4-14. We find strong evidence that polygenicity accounts for the majority of test statistic inflation in many GWAS of large sample size.

Source connections

Explore related subjects

GenomicsPreprint

Keep this discovery

BibTeXRIS

Brendan Bulik-Sullivan, Po-Ru Loh, Hilary Finucane, Stephan Ripke, Jian Yang, Schizophrenia Working Group Psychiatric Genomics Consortium, Nick Patterson, Mark J Daly, Alkes L Price, Benjamin M Neale. 2014-02-21. LD Score Regression Distinguishes Confounding from Polygenicity in Genome-Wide Association Studies. https://doi.org/10.1101/002931

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2016

Identification of genetic variants affecting vitamin D receptor binding and associations with autoimmune disease

Large numbers of statistically significant associations between sentinel SNPs and case-control status have been replicated by genome-wide association studies. Nevertheless, few underlying molecular mechanisms of complex disease are currently known. We investigated whether variation in binding of a transcription factor, the vitamin D receptor (VDR) whose activating ligand vitamin D has been proposed as a modifiable factor in multiple disorders, could explain any of these associations. VDR modifies gene expression by binding DNA as a heterodimer with the Retinoid X receptor (RXR).\n\nWe identified 43,332 genetic variants significantly associated with altered VDR binding affinity (VDR-BVs) using a high-resolution (ChIP-exo) genome-wide analysis of 27 HapMap lymphoblastoid cell lines. VDR-BVs are enriched in consensus RXR::VDR binding motifs, yet most fell outside of these motifs, implying that genetic variation often affects binding affinity only indirectly. Finally, we compared 341 VDR-BVs replicating by position in multiple individuals against background sets of variants lying within VDR-binding regions that had been matched in allele frequency and were independent with respect to linkage disequilibrium. In this stringent test, these replicated VDR-BVs were significantly (q < 0.1) and substantially (> 2-fold) enriched in genomic intervals associated with autoimmune and other diseases, including inflammatory bowel disease, Crohns disease and rheumatoid arthritis. The approachs validity is underscored by RXR::VDR motif sequence being predictive of binding strength and being evolutionarily constrained.\n\nOur findings are consistent with altered RXR::VDR binding contributing to immunity-related diseases. Replicated VDR-BVs associated with these disorders could represent causal disease risk alleles whose effect may be modifiable by vitamin D levels.

Genomics↗
bioRxiv2016

Two novel genes discovered in human mitochondrial DNA using PacBio full-length transcriptome data

In this study, we introduced a general framework to use PacBio full-length transcriptome sequencing for the investigation of the fundamental problems in mitochondrial biology, e.g. genome arrangement, heteroplasmy, RNA processing and the regulation of transcription or replication. As a result, we produced the first full-length human mitochondrial transcriptome from the MCF7 cell line based on the PacBio platform and characterized the human mitochondrial transcriptome with more comprehensive and accurate information. The most important finding was two novel lnRNAs hsa-MDL1 and hsa-MDL1AS, which are encoded by the mitochondrial D-loop regions. We propose hsa-MDL1 and hsa-MDL1AS, as the precursors of transcription initiation RNAs (tiRNAs), belong to a novel class of long non-coding RNAs (lnRNAs), which is named as long tiRNAs (ltiRNAs). Based on the mitochondrial RNA processing model, the primary tiRNAs, precursors and mature tiRNAs could be discovered to completely reveal tiRNAs from their origins to functions. The MDL1 and MDL1AS lnRNAs and their regulation mechanisms exist ubiquitously from insects to human.

Genomics↗
bioRxiv2016

Omics and bioinformatics approaches to target boar taint

In livestock species, a rapid growth in high-throughput omics data has accelerated the pace of studies that target to dissect economically important traits to provide better quality animal products to consumers. In pig industries, young boars are generally castrated to remove boar taint, a phenotypic and inheritable trait well-known by an abnormally bad smell and taste in pork meat derived from some uncastrated male pigs. Existence of porcine reference genome made possible to catalogue genome-wide QTLs, candidate genes and biomarkers in associations with boar taint and other industrially significant traits in pigs. The aim of this paper to review the contribution of bioinformatics resources and omics technology in boar taint related studies. This paper also provides concise details about state-of-the-art sequencing technology.

Genomics↗