Releases: fungenomics/CoRAL
Releases · fungenomics/CoRAL
CoRAL v4.0.3
- Fixed bug in F1 heatmap
CoRAL v4.0.2
- Seurat v5 accepted as input for reference and query
CoRAL v4.0.1
Merge pull request #161 from fungenomics/dev bug fixed on the heatmap of cawpe to dont plot it if >60000 cells (#160)
CoRAL v4.0.0
CoRAL v4.0.0 (#155) * Tom dev gene filtering (#131) * feature: feature selection method * pretrained models * added the pipeline_mode on snakemake.annotate * bug fixed in preprocess * solved issues to merge with dev * bug_fixed: now it throws an error if any tool min agreement were specified in the benchmarking config file * bug fixed: ontology rule was duplicated (#135) * Pretrained on the annotation pipeline (#140) * bug fixed: ontology rule was duplicated * implement pretrain models * bug fixed * Bug fixed scPred on pretraining models * solved issue #143 * solved issue #138 * Solved issue #141 * Solved bug * solved issue #136, now the min cell per cluster is 50 cells * feature: add minimum cells per label inside each fold, default 10 cells * Master pipeline (#144) * feature: new masterpipeline * add seed as a parameter for te preprocessing (#147) * Tom dev norm scores (#148) * add seed as a parameter for te preprocessing * clean scANVI code, extracted the scNym prediction matrix * convert correlation scores to prob * scPoli with probabilities * bug fixed: label problems for Seurat * SingleR: convert correlation to prob scores, add parallelization, add the Unkwnon prediction using the pruned process in SingleR * add comments * major changes done in the singlecellnet: add threshold as a parameter, fixed bugs related with: only using half of the reference to train, the function to call the classes has a parameter in threshold that was not working, so I create a function to call it in the same way they do it in their paper, finally, rand is remove from the probability matrix and renormalized * probability normalization, added the threshold as a parameter * clean code scPoli * bug fixed * bug fixed * Set default SVMlinear threshold to 0.5 * Changed the way that the metrics performance are calculated to take into account the Unknown categories instead of removing them * CellTypist with prob scores (#149) * Changed the way that the metrics performance are calculated to take into account the Unknown categories instead of removing them * CellTypist with probability predictions * bug fixed on celltypist * Specify some packages for avoiding problems * Issue problem in predict_scPoli.py * save cawpe scores as output (#150) * save cawpe scores as output * save cawpe scores as output * feature: norm CAWPE, re-implement CAWPE on ontology, add heatmap of CAWPE score on annotation pipeline (#154) --------- Co-authored-by: Tom <[email protected]> Co-authored-by: Tom <[email protected]>
CoRAL v3.0.2
- Bug fix (ontology)
CoRAL v3.0.1
- Bug fixes
- Updated plots in annotation notebook
CoRAL v3.0.0
- Updated structure of config file
- Weighted consensus with CAWPE: https://link.springer.com/article/10.1007/s10618-019-00638-y
- Compute consensus in higher ontology level
- New tools: Symphony, CellBlast, scPoli, scANVI, Seurat
- Downsample reference
CoRAL v2.0.3
- New figures in report + fixed colour scales
- Additional options for consensus
CoRAL v2.0.2
- fixed bug related to switch from unsure to unresolved in annotation notebook
CoRAL v2.0.1
- Bug in preprocessing script fixed