Fine mapping of GWAS and eQTL summary statistics

Sourya Bhattacharyya

La Jolla Institute for Immunology, La Jolla, CA 92037, USA

---

Implements a wrapper for the FINEMAP package (http://www.christianbenner.com/), to generate the fine mapped variants corresponding to an input GWAS or eQTL summary statistics

Prerequisites

1. Download the 1000G Genotype data (we've used hg19 as the reference genome).

   Follow the tutorial: https://cran.r-project.org/web/packages/snpsettest/vignettes/reference_1000Genomes.html.

   The path containing this 1000G genotype data needs to be provided as a configuration parameter (check the parameter GENOTYPEDIR in the configuration files, described below).

2. Install these R libraries: data.table, dplyr, GenomicRanges, gridExtra

3. Download the package LDSTORE2 (binary executable) from http://www.christianbenner.com/.

   The path of this executable needs to be provided as a configuration parameter (check the parameter ldstoreexec in the configuration files, described below)

4. Download the package FINEMAP (binary executable) from http://www.christianbenner.com/.

   The path of this executable needs to be provided as a configuration parameter (check the parameter finemapexec in the configuration files, described below)

Fine-mapping of eQTL summary statistics

First edit the configuration file configfile_eQTL.yaml.

A sample eQTL smmary statistics is provided in the file Data/Example_eQTL_data.txt for the users.

User needs to edit the configuration file according to the file / directory paths, contents of the summary statistics file.

.. Note ..

	1. In the configuration file, user needs to provide at least one of *betaCol* (column containing the beta statistics) or *ORCol* (column containing the odds ratio).

	2. Either user needs to provide the *SampleSizeCol* (column containing the sample size information) or put the integer value in the field *TotalSample* (total number of samples / donors used in the eQTL study).

	3. The parameter *geneIDCol* (column containing the gene ID, preferably in Ensemble ID format) is mandatory.

Once the configuration file is edited, user can use the script finemap_script.sh to execute the fine-mapping.

Fine-mapping of GWAS summary statistics

First edit the configuration file configfile_GWAS.yaml.

A sample GWAS smmary statistics is provided in the file Data/Example_GWAS_data.txt for the users.

User needs to edit the configuration file according to the file / directory paths, contents of the summary statistics file.

.. Note ..

	1. In the configuration file, user needs to provide at least one of *betaCol* (column containing the beta statistics) or *ORCol* (column containing the odds ratio).

	2. Either user needs to provide the *SampleSizeCol* (column containing the sample size information) or put the integer value in the field *TotalSample* (total number of samples / donors used in the eQTL study).

	3. The parameters *chrCol* (column containing the chromosome) and *posCol* (column containing the SNP position) are mandatory. 

	4. User should provide the parameter *AFCol* (column containing the allele frequency information), if available.

	5. The *OFFSET* parameter specifies the window surrounding a lead GWAS SNP used to define a GWAS locus.

Once the configuration file is edited, user can use the script finemap_script.sh to execute the fine-mapping.

Output

With respect to the specified output directory (parameter OutDir in either of these configuration files), check the folders cond (output of stepwise conditioning) and sss (output of shotgun stochastic search).

1. *FINAL_summary_credible_set.txt* : File containing the GWAS loci, credible causal sets and the corresponding SNPs.

2. *FINAL_top_snp_credible_set.txt* : All the SNPs listed in the credible causal sets. 

Contact

For any queries, please create an issue and we'll respond. Otherwise, please e-mail

Sourya Bhattacharyya: [email protected]

Ferhat Ay: [email protected]