a_johansson_2020

AJ partner project 2020

Objectives

To develop an R package for testing the scope of applicability of different GWA methodologies, esp. with respect to varying:

• threshold between rare and common variants,
• the degree of contribution (effect size and direction) contributed by rare and common variants,
• degree of population structure,
• varying amount of contributing loci.

Using on Bianca

To use the package on Bianca:

• the gwasim package is automatically built into a docker container upon every push,
• container is called gwasim-latest and is stored in quiestrho account on DockerHub,
• ssh to Rackham, do singularity pull --docker-login docker://quiestrho/gwasim:latest
• transfer the gwasim_latest.sif file into Bianca’s wharf via sftp,
• move the file from wharf to your project library,
• singularity exec gwasim_latest.sif R --vanilla < script.r to run an R script within the container

Input

The following input parameters are expected from the user:

• a VCF file with variants coming from a population under studies,
• a bed file containing a list of functional regions (e.g. gene list) along with their coordinates,
• a minor allele frequency threshold for cut-off between rare and common variants,
• the number of rare and common variants that contribute,
• distribution of effect sizes as a function of allele frequency (separately for the rare and the common alleles),
• percentage of alleles with negative effect (for both the common and the rare variants),
• parameters (mean, standard deviation) of the error term,
• type of the trait (continuous or binary, a cut-off value for the binary trait),

To add in future:

• kinship matrix for the studied population,
• fixed effects

Output

• A text file or a tibble with simulated phenotypes along with the effect of the loci used for simulation.

Possible applications:

1. To validate sensitivity and specificity of various GWA algorithms in the landscape of varying effect sizes, directionalities and mafs.