Working with linkage disequilibrium data in Ensembl
Year: 2018
Type: other
Abstract: An understanding of linkage disequilibrium (LD), the non-random association between alleles at different loci, is important when interpreting disease association data and performing population genetics studies. Since the completion of the HapMap project, the first large scale project to study genomic variation between different human populations, Ensembl has provided tools for the calculation and visualisation of LD patterns. Our LD calculation is implemented in C and follows an expectation-maximization algorithm which does not require phased data. The script calculates both r 2 and D’ values. The Ensembl Perl API provides a wrapper for the C script and implements different functionality for retrieving LD results. The API supports population-specific calculations for a given genomic region, a list of variants and for a variant and all surrounding variants in a given window size. The Ensembl browser shows both Manhattan and characteristic LD triangle plots alongside our rich gene annotation. The 1000 Genomes Project propelled the number of sequenced individuals to a new level, triggering changes in how Ensembl stores and makes genotype data available. Here, we present the changes that have been introduced to how we calculate and disseminate LD data. We moved from storing genotypes in a MySQL database to retrieving genotypes from VCF and using the htslib library for fast data retrieval. Our tools are not limited to supporting human calculations and can be used for any species with appropriate genotype data. In addition to the updated infrastructure, we will also describe the various interfaces we provide to retrieve LD results, which have been tailored to different use cases. Our REST endpoints support the same functionality as our Perl API and can be conveniently integrated into analysis pipelines and websites. REST queries are however time limited, so if the region of interest is large or variant rich we recommend using our standalone script or web tool. Our web tool presents a simple interface for configuring the LD analysis type and inputting data. Results are presented as a downloadable table with r 2 and D’ values alongside variant annotation, including predicted functional consequence and phenotype association. Lastly, we provide a Perl standalone script which hides the complexity of the Perl API and allows for easy selection of the requested LD analysis and configuration of input data.
Language: English
Subfield: Genetics
Domain: Life Sciences
Open Access status: gold
OpenAlex ID: W2899180008