Workflow from Scientific Research

cellDancer uses a DNN to predict cell-specific α , β and γ for each gene. The DNN consists of an input layer with the spliced and unspliced mRNA abundances ( u i , s i ) i = 1,2, …, n cells , two fully connected hidden layers each with 100 nodes and an output layer yielding cell-specific α , β and γ . The loss function is defined as the sum of every cell’s cosine similarity of predicted and observed velocity vectors. The DNN is iteratively optimized by minimizing the loss function.

DOI