Releases: lcrawlab/mvMAPIT
Fix LTO issues as reported by CRAN.
What's Changed
Full Changelog: v2.0.2...v2.0.3
Better control over the C++ code execution.
What has changed
- Added Cauchy combination test
cauchy_combined
including vignette that compares combination methods simulate_traits
now returns genotype matrix with causal epistatic variants named according to the trait they affect- Added progress bar and possibility to interrupt C++ routine using
RcppProgress
- Fixed "incorrect number of dimensions" error in
simulate_traits
Full Changelog: v2.0.1...v2.0.2
Package version without LTO errors due to testthat
Fixed the CRAN errors when checking with LTO enabled.
Version increment without changes for dev purposes.
v2.0.0.1 version upgrade for resubmission (#97)
First official release.
Multivariate MAPIT paper published on BiorXiv: https://www.biorxiv.org/content/10.1101/2022.11.30.518547v1
fishers_combined
andharmonic_combined
now take additional arguments in
form of string values. The first determines the name of the column of the tibble
by which to group the p-values. The second determines the name of the column
containing the p-values.- Added
Dockerfile
and vignette on mvMAPIT in Docker. MvMAPIT
was renamed tomvmapit
.simulate_phenotypes
was renamed tosimulate_traits
.- Dependencies were cleaned up, e.g. R requirement 2.10 -> 3.5, old package
dependencies removed.
Optimize memory consumption for the covariance matrices.
MM-58: refactoring should save a lot of memory (#54) [rinstall] * MM-58: refactoring should save a lot of memory * MM-58: improved davies method with C * MM-58: not sure whether tests are working
mvMAPIT only features combinatorial approach
This is the first version that only features the combinatorial approach to mvMAPIT.
Last version with Kronecker Method
This version is the last version that features the experimental kronecker method implementation.
Hybrid mode is a test mode now. Phenotype covariance contains combinatorial.
Deprecated input argument hybrid
. Now the argument test
can take the value hybrid
. Phenotype covariance now takes combinatorial
as input parameter.
p-value computation: Saddlepoint approximation when davies does not converge
The davies-method now falls back to the saddlepoint approximation as suggested in Wu et al. (2016) [1].
1.Wu, B., Guan, W. & Pankow, J. S. On efficient and accurate calculation of significance p-values for sequence kernel association testing of variant set. Ann Hum Genet 80, 123–135 (2016).