Dual deep encoder-based framework for single-cell RNA-seq data clustering

Single-cell RNA-seq data clustering plays an important role in understanding the heterogeneity of different tissues at the cellular level by differentiating various cell types, which ultimately helps disease prediction. Analyzing scRNA-seq data is challenging because of its high-dimensional, complex, and noisy nature. Recently, deep learning techniques boosted this analysis to a higher level, outperforming traditional statistical and machine learning techniques. Here, we propose a deep learningbased clustering model utilizing two parallel encoders to obtain latent embedding space, considering the manifold structure of the scRNA-seq data. The proposed method registered higher performance in different metrics than several state-of-the-art methods.