Rxivist logo

Efficient Integrative Multi-SNP Association Analysis using Deterministic Approximation of Posteriors

By Xiaoquan Wen, Yeji Lee, Francesca Luca, Roger Pique-Regi

Posted 09 Sep 2015
bioRxiv DOI: 10.1101/026450 (published DOI: 10.1016/j.ajhg.2016.03.029)

With the increasing availability of functional genomic data, incorporating genomic annotations into genetic association analysis has become a standard procedure. However, the existing methods often lack rigor and/or computational efficiency and consequently do not maximize the utility of functional annotations. In this paper, we propose a rigorous inference procedure to perform integrative association analysis incorporating genomic annotations for both traditional GWAS and emerging molecular QTL mapping studies. In particular, we propose an algorithm, named ``Deterministic Approximation of Posteriors" (DAP), which enables highly efficient and accurate joint enrichment analysis and identification of multiple causal variants. We use a series of simulation studies to highlight the power and computational efficiency of our proposed approach and further demonstrate it by analyzing the cross-population eQTL data from the GEUVADIS project and the multi-tissue eQTL data from the GTEx project. In particular, we find that genetic variants predicted to disrupt transcription factor binding sites are enriched in cis-eQTLs across all tissues. Moreover, the enrichment estimates obtained across the tissues are correlated with the cell types for which the annotations are derived. The software package implementing the computational approach (including a detailed tutorial) is freely available at https://github.com/xqwen/dap.

Download data

  • Downloaded 1,033 times
  • Download rankings, all-time:
    • Site-wide: 20,026
    • In genetics: 964
  • Year to date:
    • Site-wide: 51,753
  • Since beginning of last month:
    • Site-wide: 52,263

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide


Sign up for the Rxivist weekly newsletter! (Click here for more details.)