The DigitalDLSorterDNN object stores all the information related to Deep Neural Network models. It contains the trained model, the training history and the results of prediction on test data. After running calculateEvalMetrics, it is possible to find the performance evaluation of the model on test data (see ?calculateEvalMetrics for details).


The steps related to Deep Learning are carried out using the keras package which uses the R6 classes system. If you want to save the object as an RDS file, digitalDLSorteR provides a saveRDS generic function that transforms the model stored as an R6 object into a native valid R object. Specifically, the model is converted into a list with the architecture of the network and the weights learned during training. That is the minimum information needed to use the model as predictor. If you want to keep the optimizer state, see ?saveTrainedModelAsH5. If you want to store DigitalDLSorter object on disk as an RDA file, see ?preparingToSave.



Trained Deep Neural Network. This slot can contain an R6 keras.engine.sequential.Sequential object or a list with two elements: the architecture of the model and the resulting weights after training.


List with the evolution of the selected metrics during training.


Performance of the model on test data.


Deconvolution results on test data. Columns are cell types, rows are samples and each entry corresponds to the proportion of this cell type in this sample.


Vector with cell types to deconvolute.


Vector with the features used during training. These features will be used in subsequent predictions (the nomenclature used in new bulk RNA-Seq samples must be the same).


Performance of the model on each sample of test data compared to known cell proportions. This slot is used after calculateEvalMetrics (see ?calculateEvalMetrics for more details).