You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Suhas Srinivasan edited this page Dec 26, 2018
·
5 revisions
Deep unsupervised single-cell clustering (DUSC), is a hybrid approach that integrates feature-selection based on a deep learning architecture with a clustering algorithm, to find a compressed and informative representation of single-cell transcriptomic data.
DUSC is resilient to the inherent biological and technical noises in single-cell experiments and is used to generate cell clusters corresponding to cell types and sub-types.
We also introduce the denoising autoencoder with neuronal approximation (DAWN), which can efficiently estimate the number of latent features required for each dataset.