Data Resources
Computational Analysis Tools
Standards
Data Hubs
| Name | mVAEN |
|---|---|
| Type | Health and Disease Detection |
| Developers | Peilin Jia |
| Description | mVAEN is a deep generative method to predict drug response based on baseline expression data. It utilizes variational autoencoders (VAE) with a mask layer and compresses transcriptome data into latent vectors in a low-dimension space. The newly introduced mask layer incorporates a prior gene-pathway mapping and, thus, it provides interpretable gene embeddings. The resultant gene beddings are subsequently used to train Elastic Net (EN) models for drug response prediction. Because mVAEN can be trained using a baseline gene expression profile, it can be applied in a wide range of clinical settings. |
| Downlaod | https://bioinfo.uth.edu/drvaen/ |
| Article | https://academic.oup.com/bib/article-abstract/24/3/bbad098/7085476 |
1 Beichen West Road, Chaoyang District Beijing 100101, China | 86-10-84097216
© China National Center for Bioinformation 2025, 京ICP备 10050270号-13
