coloc.compute: Colocalization computation
In tobyjohnson/gtx: Genetics ToolboX

Description Usage Arguments Details Value Author(s)

Colocalization computation using summary statistics

1
2
3

coloc.compute(data, priorsd1 = 1, priorsd2 = 1, priorc1 = 1e-04,
  priorc2 = 1e-04, priorc12 = 1e-05, join_type = "inner",
  summary_only = FALSE)

`data`	Data frame with one row for each variant, and columns beta1, se1, beta2, se2
`priorsd1`	Standard deviation of prior on beta1
`priorsd2`	Standard deviation of prior on beta2
`priorc1`	Prior on variant being causal for trait 1
`priorc2`	Prior on variant being causal for trait 2
`priorc12`	Prior on variant being causal for traits 1 and 2
`join_type`	How to handle missing summary statistics
`summary_only`	Whether to return only the colocalization results

Computes Bayesian posterior probabilities of colocalization versus alternative models, based on the approach of Giambartolomei et al. (2014) but with a more efficient and more flexible implementation. The original approach is also extended to compute a model averaged point estimate of the causal effect of each trait on the other.

The parameter join_type controls how variants with missing summary statistics are handled. inner subsets to variants with non-missing statistics for both analyses. outer keeps substitutes Bayes factors of zero for all variants with missing statistics (i.e. assumes they have much less strong associations than other variants). left subsets to variants with non-missing beta1 and se1 analysis 1, and substitutes Bayes factors of zero for variants with missing beta2 or se2. right does the opposite.

The input data frame with additional columns added, and colocalization probabilities as attributes.

Toby Johnson Toby.x.Johnson@gsk.com

tobyjohnson/gtx documentation built on Aug. 30, 2019, 8:07 p.m.