Simulated Summary Statistics with LD Reference Mismatch.
Source:R/example_dataset.R
rss_mismatch_cases.RdFive simulated regions, each with one causal variant and
121 variants, used to illustrate how susie_rss handles a
mismatch between the summary statistics and the LD reference. The
first four cases simulate a phenotype from GTEx genotypes (n = 838);
the last is a meta-analysis of an African-ancestry and a
European-ancestry cohort.
Format
rss_mismatch_cases is a list of five cases. Each case is
a list with the inputs to susie_rss:
- z
z-scores.
- R
LD matrix; for
multi_panel, a named list of LD matrices, one per reference panel.- X
For
ld_mismatch_sketchonly, a genotype sketch (factor matrix) of the reference panel, used instead ofR. This case also storesR, a copy ofR_insample.- n
Sample size of the summary statistics.
- B
Reference panel sample size, or the sketch dimension (number of rows of
X), passed asR_finite; one value per panel formulti_panel. Absent forallele_flip, whose LD is in-sample.- causal
Index of the true causal variant.
The cases are:
- allele_flip
In-sample LD, with the sign of the z-score reversed at 1% of the variants. Also contains
z_clean, the z-scores before flipping, andflip_id, the indices of the flipped variants.- ld_mismatch
LD from 489 European individuals in 1000 Genomes.
- ld_mismatch_sketch
The reference panel given as a 500-row genotype sketch
Xof the ADSP panel.- ser_fallback
LD from 489 European individuals in 1000 Genomes, a region where
SER_fallbackisTRUE.- multi_panel
Meta-analysis of 400 African-ancestry and 400 European-ancestry samples, with AFR and EUR panel LD in
R. Also containsprop, the true proportion of samples from each ancestry.
ld_mismatch, ld_mismatch_sketch and ser_fallback
also contain z_insample and R_insample, the z-scores and
the in-sample LD on the same variants.