v0.1.2

Changed

• Order of features in feature matrix for prediction is ensured to be the same as during training of the model
• Cellular contexts labelled testing/validation are removed from binding pattern computation

Fixed

• Normalization of NA-containing columns

v0.1.1

Changed

Fixed

• Ensuring consistent ordering and generic naming of tf-cofactors across features and in col- & metadata.
• trainTfModel(): Avoid breaking when training contexts contain zero positive labels by removing them from training
• tfFeatures(): Avoid breaking when co-binding features are requested but no tf-cofactors provided

First initial package version

• In terms of code and functionality identical to version 0.0.1.1001 as specified earlier in the R package description file