Nothing
confint() with deconv objects to allow calculation of
confidence intervals.max_similarity() for calculating maximum similarity between cell types.deconvolute() [experimental].$opt element to store call arguments in deconvolute().plot_comp() and plot_path() with option to show cell percentages.deconvolute() to help with issue #20.tune_deconv().plot_tunevar() to compare MSE, bias^2 and variance.kappa() to compute the condition number of the spillover
matrix.plot_path() to show how the coefficient paths vary with compensation for
either a single bulk sample or a cell subclass.plot_pred() as an alternative to plot_set() for viewing predicted vs
observed results from simulations in a single correlation plot.zero_fraction argument to generate_samples().nest_outputdeconvolute().lambda to deconvolute() [experimental].resvar metric to tune_deconv() to examine residual variance of bulk
gene expression.deconvolute() no longer returns se and hat in the subclass list
element. Standard errors can be obtained with the new function se().cores argument has been removed from deconvolute()mclapply().
The problem is isolated to R 4.5.2 arm64 for macOS on M Macs; or any version of
R for macOS including intel, if vecLib BLAS is being used via a symlink. The
solution is to use cores = 1 with deconvolute() and tune_deconv() whenever
vecLib BLAS is in use.pbmcapply::pbmclapply() as this caused problems with R 4.5.2
for macOS arm64 (Apple M* Macs) even with cores=1. This fixes indefinite
hanging in tune_deconv() associated with vecLib BLAS.log argument in deconvolute() to logged_bulk. NB. this is a
change of logic.residuals.deconv() to allow recalculation of full residuals matrix.mergeMarkers() if cellMarkers object has no cell grouping vector.cat() to message()tune_deconv()Any scripts or data that you put into this service are public.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.