General purpose scripts
Given a table with meta-analysis results for independent hits, create an extended table containing information about known signals in a specified bp window around hits.
$ cojo-prepare.sh -i <input table> -o <output table> -k <known signals> -w <bp window>
- input table: a tab separated table containing meta-analysis results for independent hits; should contain following columns:
- column 1: panel
- column 2: protein
- column 5: UniProt ID of the protein
- column 6: chromosome of the hit
- column 8: position of the hit
- column 10: effect allele
- column 11: other allele
- column 12: frequency of the effect allele
- column 16: beta
- column 17: standard error
- column 18: p-value
- column 24: number of missing samples in meta-analysis of the current hit
- output table: a tab separated table containing
- record ID of the form: PANEL_PROTEIN_CHR_POS, where CHR and POS refer to the genomic position of the independent hit from the input table
- CHR: chromosome of the signal (either input hit, or conditioning variant)
- POS: position of the signal (either input hit, or conditioning variant)
- A1: effect allele
- A2: other allele
- freq: allele frequency of the effect allele
- beta: effect size
- SE: standard error of the effect size
- p-value
- N: number of samples analyzed <-- redundant
- sample names: comma-separated list of samples used in the meta-analysis
- known signals: a BED file containing information about known associations; should at least contain the following fields:
- 1: chromosome of the known signal
- 2: pos-1, where pos is the position of the known signal
- 3: position of the known signal
- 5: UniProt ID of the associated protein
- optional bp window restricting conditioning known signals (default: 1Mbp)
Given an output table from the cojo-prepare.sh
step, perform GCTA cojo-cond
testing of the independent hits conditioned on known signals.
$ cojo-wrapper.sh -f <PLINK bfile prefix> -o <output directory> -i <input m/a table>
- PLINK bfile prefix: prefix filename of bed/bim/fam PLINK files
- output directory: directory with the intermediate PLINK files and result GCTA files, also contains
cojo-wrapper.log
andcojo-wrapper.err
log files.cojo-wrapper.err
contains information about failed GCTA runs - input m/a table: output table from the
cojo-prepare.sh
run
A simple wrapper for GCTA cojo-slct
analysis, intended to be used when cojo-wrapper.sh
fails because of collinearity between conditioning variants.
$ cojo-slct.sh -i <signal ID> -n <number of samples> -f <PLINK bfile prefix>
- signal ID: ID of the failed signal of the form PANEL_PROTEIN_CHR_POS (these IDs form the first column of
cojo-wrapper.err
error file of the previouscojo-wrapper.sh
run) - number of samples: number of samples used in the m/a analysis
- PLINK bfile prefix: prefix filename of bed/bim/fam PLINK files