Elucidation of cell subpopulations at high resolution is a key and challenging goal of single-cell ribonucleic acid (RNA) sequencing (scRNA-seq) data analysis. Although unsupervised clustering methods have been proposed for de novo identification of cell populations, their performance and robustness suffer from the high variability, low capture efficiency and high dropout rates which are characteristic of scRNA-seq experiments. Here, we present a novel unsupervised method for Single-cell Clustering by Enhancing Network Affinity (SCENA), which mainly employed three strategies: selecting multiple gene sets, enhancing local affinity among cells and clustering of consensus matrices. Large-scale validations on 13 real scRNA-seq datasets show that SCENA has high accuracy in detecting cell populations and is robust against dropout noise. When we applied SCENA to large-scale scRNA-seq data of mouse brain cells, known cell types were successfully detected, and novel cell types of interneurons were identified with differential expression of gamma-aminobutyric acid receptor subunits and transporters. SCENA is equipped with CPU + GPU (Central Processing Units + Graphics Processing Units) heterogeneous parallel computing to achieve high running speed. The high performance and running speed of SCENA combine into a new and efficient platform for biological discoveries in clustering analysis of large and diverse scRNA-seq datasets.

Single-cell RNA sequencing (scRNA-seq) is a widely used technique for characterizing individual cells and studying gene expression at the single-cell level. Clustering plays a vital role in grouping similar cells together for various downstream analyses. However, the high sparsity and dimensionality of large scRNA-seq data pose challenges to clustering performance. Although several deep learning-based clustering algorithms have been proposed, most existing clustering methods have limitations in capturing the precise distribution types of the data or fully utilizing the relationships between cells, leaving a considerable scope for improving the clustering performance, particularly in detecting rare cell populations from large scRNA-seq data. We introduce DeepScena, a novel single-cell hierarchical clustering tool that fully incorporates nonlinear dimension reduction, negative binomial-based convolutional autoencoder for data fitting, and a self-supervision model for cell similarity enhancement. In comprehensive evaluation using multiple large-scale scRNA-seq datasets, DeepScena consistently outperformed seven popular clustering tools in terms of accuracy. Notably, DeepScena exhibits high proficiency in identifying rare cell populations within large datasets that contain large numbers of clusters. When applied to scRNA-seq data of multiple myeloma cells, DeepScena successfully identified not only previously labeled large cell types but also subpopulations in CD14 monocytes, T cells and natural killer cells, respectively.

Too many input features in applications may lead to over-fitting and reduce the performance of the learning algorithm. Moreover, in most cases, each feature containing different information content has different effects on the prediction target. Therefore, a feature selection method for calculating the importance of each feature, called WKNNGAFS, is proposed in this paper. In this method, the genetic algorithm (GA) is adopted to search the optimal weight vector, the value of the ith component of which corresponds to the contribution degree of the ith feature to the classification from a global perspective. Besides, weighted K-nearest neighbors algorithm (WKNN), which takes both the different contributions of nearest neighbors and the different classification ability of each feature into account, is used to determine the target label. To evaluate the effectiveness of the proposed method, nine existing feature selection methods are compared with it on 13 real datasets, including 6 high dimensional microarray datasets. Experimental results demonstrate the method is more effective and can improve classification performance.

Although cis-regulatory binding sites (CRBSs) are at least as important as the coding sequences in a genome, our general understanding of them in most sequenced genomes is very limited due to the lack of efficient and accurate experimental and computational methods for their characterization, which has largely hindered our understanding of many important biological processes. In this article, we describe a novel algorithm for genome-wide de novo prediction of CRBSs with high accuracy. We designed our algorithm to circumvent three identified difficulties for CRBS prediction using comparative genomics principles based on a new method for the selection of reference genomes, a new metric for measuring the similarity of CRBSs, and a new graph clustering procedure. When operon structures are correctly predicted, our algorithm can predict 81% of known individual binding sites belonging to 94% of known cis-regulatory motifs in the Escherichia coli K12 genome, while achieving high prediction specificity. Our algorithm has also achieved similar prediction accuracy in the Bacillus subtilis genome, suggesting that it is very robust, and thus can be applied to any other sequenced prokaryotic genome. When compared with the prior state-of-the-art algorithms, our algorithm outperforms them in both prediction sensitivity and specificity.

Motivation: The availability of numerous ChIP-seq datasets for transcription factors (TF) has provided an unprecedented opportunity to identify all TF binding sites in genomes. However, the progress has been hindered by the lack of a highly efficient and accurate tool to find not only the target motifs, but also cooperative motifs in very big datasets. Results: We herein present an ultrafast and accurate motif-finding algorithm, ProSampler, based on a novel numeration method and Gibbs sampler. ProSampler runs orders of magnitude faster than the fastest existing tools while often more accurately identifying motifs of both the target TFs and cooperators. Thus, ProSampler can greatly facilitate the efforts to identify the entire cis-regulatory code in genomes. Availability and implementation: Source code and binaries are freely available for download at https://github.com/zhengchangsulab/prosampler. It was implemented in C++ and supported on Linux, macOS and MS Windows platforms. Supplementary information: Supplementary materials are available at Bioinformatics online.