HetEnc

HetEnc is a deep learning-based tool to integrate and analyze multi-platform gene expression data for single or multi-task classification problems in biological research. The tool consists of two main modules:

1. An unsupervised feature representation module that constructs three different encoding networks to represent the original gene expression data using high-level abstracted features.

2. A supervised neural network module with a six-layer fully connected feed-forward neural network, which is trained using the abstracted features for each targeted endpoint.

HetEnc's key advantage is its ability to handle multi-platform data during feature abstraction and model training while requiring only single-platform data for prediction. It reduces gene expression profiling costs for new samples and enables broader application of the trained model.