extract_instruments() and check_phenotypes() now harmonise with TwoSampleMR action = 1, assuming all alleles are on the forward strand as they are in OpenGWAS. Previously the default action = 2 inferred the strand of palindromic SNPs from their allele frequencies, which differ between populations, so palindromic SNPs whose allele frequencies were on different sides of 0.5 in the reference population had their beta and eaf wrongly flipped, and palindromic SNPs with intermediate allele frequencies were dropped. extract_instruments() gains a harmonise_strictness argument.
Fix the FDR adjustment in estimate_instrument_heterogeneity_per_variant() and mrgxe_plot_variant(). It was applied separately to each SNP, so Qfdr equalled Qpval and the default SNP selections in mrgxe() and mrgxe_plot_variant() used unadjusted p-values. mrgxe() now returns an ungrouped tibble.
Bump the required version of ieugwasr
CAMeRa 0.1.3
Fix CAMERA_local so it can be initialised and run: add the missing radius field, store ld_ref, call the class’s own methods, fix the exposure/outcome subsetting, stop mc.cores being passed as pthresh, use mc.cores (default 1, so it works on Windows) in all parallel calls, use the minmaf setting when reading files, and declare its dependencies (data.table, GenomicRanges, IRanges, parallel) in Suggests (thanks @Mercy-Kimani)
CAMeRa 0.1.2
Additional amends to the tutorial
Show loading of the example data and state when it was created
Explain how to install genetics.binaRies or supply a plink executable
Remove the unused x$clone() and clarify how make_outcome_data() and harmonise() update the object
Add an overview of the CAMERA methods with links to their documentation
Regenerate the example data used in the tutorial from OpenGWAS, using an ieugwasr version which corrects the signs of the beta and eaf of the ukb-e datasets returned by associations(), and update the tutorial text to match the new results
Fix swapping of the effect and other alleles in extract_instrument_regions() when a dataset’s effect allele differs from the first dataset’s
Expand the documentation of the instrument_heterogeneity(), estimate_instrument_specificity() and cross_estimate() methods, including their output columns and the null hypotheses of their p-values