Changelog • dominoSignal

dominoSignal v1.0.3

Added functions for calculation of mean gene expression among components of a complex using purrr functions

Restored _pkgdown.yml and index scripts that specify site building parameters for https://FertigLab.github.io/dominoSignal to improve formatting.

Resolved issue where if create_domino was run with complexes=TRUE and no complexes were found to have active signaling, the full signaling matrix would be replaced by a NULL value.

Disabled exact p-value computation for the correlation test between receptor expression and features to prevent repeated warning messages due to tied ranks during Spearman correlation calculation

Updated non-functional links to functional URLs
All vignettes explicitly state the seed used when executing their code
The dominoSignal Object vignette states the purpose of the code used to download and import data from a BioFileCache to demonstrate applications of the domino objects on a real data object that is too large to include inside the package

Data used for examples and unit tests is now stored in the package’s data/ directory.
All usage of the triple colon operator to access internal data from sysdata have been replaced with usage of data()
Linkage summaries for demonstrating differential linkage testing and plotting are generated within unit test scripts rather than being stored within the package.
Examples are run with echo = FALSE to cut down on lines of code printed in example pages.
Examples for create_domino are run with verbose = FALSE
Fixed example of changing colors for circos_ligand_receptor
Fixed example of boolean representation of cor_heatmap

Update to vignettes presenting application of the DominoSignal pipeline on data formatted as a SingleCellExperiment object
Implemented caching of example data by BiocCache to meet package size limits
Removal of deprecated scripts for running SCENIC. Tutorials for running SCENIC are still present in vignettes
Corrected BiocCheck notes pertaining to coding practices including paste in conditional statements, functions with dontrun examples, usage of seq_len or seq_along in place of seq, and usage of vapply in place of sapply

Uniform formats for inputs of receptor - ligand interaction databases, transcription factor activity features, and regulon gene lists for operability with alternative databases and transcription factor activation inference methods
Helper functions for reformatting pySCENIC outputs and CellPhoneDB database files to domino-readable uniform formats
Assessment of transcription factor linkage with receptors that function as a heteromeric complex based on correlation between transcription factor activity and all receptor component genes
Assessment of complex ligand expression as the mean of component gene expression for plotting functions
Minimum threshold for the percentage of cells in a cluster expressing a receptor gene for the receptor to be called active within the cluster
Additional linkage slots for active receptors in each cluster, transcription factor - receptor linkages for each cluster, and incoming ligands for active receptors on each cluster

Chord plot of ligand expression targeting a specified receptor where chord widths correspond to the quantity of ligand expression by each cell cluster
Signaling networks showing only outgoing signaling from specified cell clusters
Gene networks between two cell clusters

Added host option for gene ortholog conversions using biomaRt for access to maintained mirrors
Transcription factor - target linkages are now properly stored so that receptors in a transcription factor’s regulon are excluded from linkage
Ligand nodes sizes in gene networks correspond to quantity of ligand expression
create_domino() can be run without providing a regulon list
References to the host GitHub repository have been updated to Elisseeff-Lab