LD matrices from the White British cohort in the UK Biobank in Zarr format

<p>This dataset contains the Linkage Disequilibrium (LD) matrices that were used in the analyses described in the manuscript:</p> <p><strong>Fast and Accurate Bayesian Polygenic Risk Modeling with Variational Inference</strong><br> Shadi Zabad, Simon Gravel, Yue Li<br> McGill University</p> <p>LD matrices record the SNP-by-SNP correlations in a given sample of individuals from a general population. In this case, we threshold the matrices so that we only record the correlations between SNPs that are at most 3 centi Morgan apart. These matrices record the SNP correlations in a random sample of 50,000 individuals&nbsp;from the White British cohort in the UK Biobank dataset. There is one matrix per autosomal chromosome (chr_1, chr_2, ..., chr_22). The matrices are stored in <a href="https://zarr.readthedocs.io/en/stable/">Zarr</a> format, a chunked on-disk array storage format that allows for multi-threaded read and write access.</p> <p>To access these matrices, consult the codebase of <a href="https://github.com/shz9/magenpy"><strong>magenpy</strong></a>, our custom python package with special data structures for processing these LD matrices.</p> <p>UPDATE (03/09/2022): We updated the matrices to add the reference allele attribute (A2) and we also now have one tar archive per chromosome.<br> &nbsp;</p>