Analyses summary statistics from genome-wide association studies of cohorts with mixed ancestry, against an external reference panel of 32,953 samples drawn from 29 populations. The panel is read directly off disk by position, so it is never loaded into memory and is not distributed with the package.
Details
The entry points fall into four groups:
- Imputation of unmeasured variants
distfor a single population or superpopulation,distmixfor an arbitrary ancestry mixture.- Gene-level association
jepegandjepegmix, which aggregate functional annotations into a single statistic per gene.- Region-level association
- Ancestry and linkage disequilibrium
afmixandzmixestimate mixing proportions from the study's own allele frequencies or Z-scores;computeLDreturns the ancestry-informed LD matrix the other functions use internally.
Functions whose name ends in mix take a pop_wgt_df of mixing
proportions and are read positionally: the first column must name the
population and the second must carry its weight. See
data(PGC2_SCZ_ANC_Prop) for the expected shape.
References
Lee D, Williamson VS, Bigdeli TB, Riley BP, Fanous AH, Vladimirov VI, Bacanu SA (2015). DIST: direct imputation of summary statistics for unmeasured SNPs. Bioinformatics 31(19):3099-3104. doi:10.1093/bioinformatics/btv348