Network from Scientific Research

Method overview illustrating the neural network architecture used for Training and Imputation. CMImpute’s CVAE framework takes as input a matrix of individual observed samples with corresponding species and tissue labels. During Training , the CVAE learns methylation patterns from the three categories of training data. Once trained, Imputation can occur. The CVAE uses the learned parameters to impute species-tissue combination mean samples of the missing target species-tissue combinations. In the example illustrated, CMImpute imputes the missing horse tissues. For visualization purposes, X, y, X' , and X impute are shown transposed (Methods)

DOI