NEWS

bakR 1.0.1 (2024-01-13)

Updated Stan array syntax to be compatible with current rstan version
Hack to address Hmisc binning edge case (see https://github.com/harrelfe/Hmisc/issues/169).

Functions for visualizing (VisualizeDropout), quantifying (QuantifyDropout), and correcting (CorrectDropout) metabolic label-induced dropout of RNA during library preparation have been added.
New simulation function (simulate_relative_bakRData) which better captures the relative nature of RNA-seq and can accurately simulate dropout.
New experimental function (DissectMechanism) for determining how likely that any observed differential expression is driven by transcriptional or post-transcriptional regulation. DissectMechanism is a rewrite and extension of the previously developed NSSHeat2 function, which itself was an improvement of the now deprecated NSSHeat.
Can now provide fraction new estimates (e.g., from a tool like GRAND-SLAM) as input to bakR. GRAND-SLAM input functionality is further supported by the new GSprocess function that will facilitate converting from GRAND-SLAM output to bakR input.
FnPCA has been deprecated in favor of FnPCA2 which accepts input differently and fixes some bugs.
Read count filtering now includes two filters. One read count that all samples must pass, and one that only all replicates in a single Exp_ID need to pass. This facilitates identifying large increases or decreaes in expression.
Several new vignettes to discuss much of the new functionality discussed above.
Several small bug fixes

Small edit to configuration files that address compilation issues that can arise on some systems. Deals with "file too big" errors" during package installation from source.

Implemented long read sequencing data analysis strategy. Run bakRFit() with Long = TRUE to use k-means clustering (k = 2) for mutation rate and fraction new estimation. Need to have Ckmeans.1d.dp package installed to do this (not installed during bakR installation).

Fixed bug in reliableFeatures. high_p was supposed to be the maximum allowable mutation rate (# of mutations/# of Ts) in reads from -s4U controls, but was instead the maximum allowable average number of T-to-C mutations in reads from these controls.

Fixed bug in cBprocess that didn't properly check that features of interest provided by the FOI argument were valid.

Added QC_check(), a function to perform quality control analysis with bakRFit objects. Looks for any problems in your data that will impair bakR's performance, generates a number of diagnostic visualizations, and makes suggestions about what to do next.
Fixed plot coloring bug in plotMA() and plotVolcano().
Fixed bug that led to problems when the number of -s4U replicates > +s4U replicates in one or more Exp_IDs
Implemented improved U-content adjustment for MCMC implementation. Also impacts accuracy of StanRateEst = TRUE mutation rate estimation strategy.
Increased default number of features to use for StanRateEst mutation rate estimation strategy.
Improved scaling of NSSHeat() output matrix columns.
Created a new function NSSHeat2() that implements a different mechanism scoring function than NSSHeat().