This function is a wrapper around CARMA::CARMA() that takes a dataframe containing variants at a locus and performs Bayesian finemapping based on their signed z-scores and an LD reference panel. The function requires a plink-formatted LD reference panel (bfile), which will be used to generate a signed LD matrix at the locus. CARMA was described in Yang et al. (Nature Genetics 2023; https://doi.org/10.1038/s41588-023-01392-0)
Usage
run_carma(
df,
snp_col,
z_col,
effect_allele_col,
outlier_switch = TRUE,
bfile,
threads = 1,
memory = 16000,
plink_bin,
...
)Arguments
- df
Dataframe containing variants at a locus for finemapping
- snp_col
Name of column containing SNP identifiers
- z_col
Name of column containing signed Z-scores (relative to effect allele)
- effect_allele_col
Name of column containing effect alleles
- outlier_switch
Whether to remove outliers when performing finemapping. Argument is passed to the
outlier.switchargument ofCARMA::CARMA()- bfile
Path to plink
bfileof reference panel that will be used to extract LD- threads
Number of threads (default = 1)
- memory
Memory limit (default = 16000 MB)
- plink_bin
Path to plink executable
- ...
additional arguments passsed to
CARMA::CARMA()
Value
A dataframe containing the input dataframe, and additional columns denoting which credible set (CS) each variant belongs to, as well as the posterior inclusion probability (PIP), and an ld_error column noting whether there were problems generating the LD matrix that limited fine-mapping.
See also
Other finemapping:
calc_credset()