Skip to contents

Changelog

Source: NEWS.md

mfsusieR 0.0.1 (2026-04-25)

First release.

mfsusieR is now official implementation of the fSuSiE and mfSuSiE models on the susieR backbone. It unifies the previously separate fsusieR and mvfsusieR codebases under one package; both methods, as described in the corresponding manuscripts, are now distributed through mfsusieR.

What is in this release:

  • mfsusie() for multi-outcome functional fine-mapping; each outcome can be a scalar or a curve sampled at a grid of positions.
  • fsusie() for single-outcome functional fine-mapping (M = 1, T_1 > 1) with the conventional (Y, X, pos, ...) argument order.
  • S3 methods: predict.mfsusie(), coef.mfsusie(), fitted.mfsusie(), summary.mfsusie(), print.mfsusie(), and plot.mfsusie().
  • mf_post_smooth(fit, method = c("TI", "scalewise", "HMM", "smash"), overwrite_previous = FALSE) for posterior- smoothing of effect curves with credible bands. "TI" is the default (translation-invariant wavelet denoising via cycle spinning); "HMM" additionally returns per-position lfsr; "smash" delegates to smashr::smash.gaus (Suggests). The "scalewise" pointwise variance is computed via the squared inverse-DWT matrix (Var(pos[t]) = sum_k W^T_{tk}^2 * var_w[k]).
  • Each mf_post_smooth() call adds an entry to fit$smoothed[[method]] rather than overwriting top-level slots, so multiple smoothers coexist on the same fit. With overwrite_previous = FALSE (default), re-applying the same smoother errors instead of clobbering. coef(fit) returns the raw inverse-DWT curves; pass coef(fit, smooth_method = "<name>") for a smoothed version. mfsusie_plot() picks a smoother by priority TI > smash > HMM > scalewise when several are present and emits a hint listing the alternatives; pass smooth_method = "<name>" to plot a specific one.
  • mf_adjust_for_covariates(Y, Z, X = NULL, method = c("wavelet_eb", "ols")) for pre-fit covariate adjustment of a functional response. method = "wavelet_eb" fits a wavelet-domain empirical-Bayes regression of Y on Z; method = "ols" is the closed-form linear regression covariate adjustment using ordinary least square method.
  • mfsusie_plot() and mfsusie_plot_lfsr() for visualizing PIPs, credible-set membership, per-CS effect curves with optional credible bands, and per-CS local false sign rates.
  • Bundled simulated example datasets: dnam_example, rnaseq_example, multiomic_example, gtex_example. Each ships with a build script under data-raw/.