SignacX_2.2.0

Likely a CRAN-stable release; pending review at CRAN.

signac_2.1.1

Introduced SignacFast, a faster method for cell type classification using pre-trained models as opposed to running training using intersected features between model and target data. Added vignettes, as well as a "ModelGenerator" function for building machine learning models.

signac_2.1.0

Signac version 2.1.0

• Allowed Signac to read the default assay from Seurat, performing analysis on RNA or SCT assay.
• Included additional vignettes in the GitHub repository.

signac_2.0.10

Patched an issue related to parallelization seen in one data set on an AWS instance. Unclear what caused the problem; some cores returned an error from neural network.

signac_2.0.9

Fixed bug: Modified FindAllMarkers to have 500 maximum cells per ident. This might help that function from hanging.

signac_2.0.8

Bug fixes

1. Changed Signac solo to use 100 neural networks.
2. Switched from rjson to RJSONIO package to prevent I/O error.
3. Added verbose feature in SignacLearn

signac_2.0.7

Added set.seed functionality in parallel environment in Signac function to allow for reproducible results for manuscript

signac_2.0.6

Implemented pathway enrichment functions, changed smoothing protocol to not smooth nuanced immune phenotypes.

signac_2.0.5

Signac v2.0.5

1. Fixed issue where user runs Signac on a Seurat object, does not want to run imputation. Previously, Signac hid Louvain clustering computation inside an "if" statement that returned an error if the user did not run imputation. This problem was fixed and requires no further study.
2. Fixed an issue with loading "hdf5" networks. There was a dimensionality issue. Although the networks load now and appear to be fine (e.g., the diagonal of the adjacency matrix is 1, indicating proper indices mapping from the hdf5 files), this might require some future investigation into what caused the error.
3. The package used for graphical LASSO is now glasso, due to CRAN no longer maintaining QUIC.

signac_2.0.4

1. Fixed bugs to allow for seamless integration of Seurat.
2. Allowed functions to define novel cell type populations and generate color tracks for them in SPRING.
3. Included a vignette on SPRING and Seurat data processing with Signac.