P-CSN: single-cell RNA sequencing data analysis by partial cell-specific network

Abstract

Although many single-cell computational methods proposed use gene expression as input, recent studies show that replacing ‘unstable’ gene expression with ‘stable’ gene–gene associations can greatly improve the performance of downstream analysis. To obtain accurate gene–gene associations, conditional cell-specific network method (c-CSN) filters out the indirect associations of cell-specific network method (CSN) based on the conditional independence of statistics. However, when there are strong connections in networks, the c-CSN suffers from false negative problem in network construction. To overcome this problem, a new partial cell-specific network method (p-CSN) based on the partial independence of statistics is proposed in this paper, which eliminates the singularity of the c-CSN by implicitly including direct associations among estimated variables. Based on the p-CSN, single-cell network entropy (scNEntropy) is further proposed to quantify cell state. The superiorities of our method are verified on several datasets. (i) Compared with traditional gene regulatory network construction methods, the p-CSN constructs partial cell-specific networks, namely, one cell to one network. (ii) When there are strong connections in networks, the p-CSN reduces the false negative probability of the c-CSN. (iii) The input of more accurate gene–gene associations further optimizes the performance of downstream analyses. (iv) The scNEntropy effectively quantifies cell state and reconstructs cell pseudo-time.

Building and analyzing metacells in single-cell genomics data

“…Many tools, such as MetaCell (Baran et al, 2019 ), MC2 (Ben-Kiki et al, 2022 ), SuperCell (Bilous et al, 2022 ) and SEACells (Persad et al, 2023 ), are available to build metacells. Metacells have been effectively employed in various studies for multiple types of downstream analyses (Li et al, 2019 ; Giladi et al, 2018 ; Cohen et al, 2018 ; Bornstein et al, 2018 ; Zheng et al, 2021b ; Cohen et al, 2022 ; Liu et al, 2021b ; Nguyen et al, 2020 ; Magaletta et al, 2022 ; Wang et al, 2023 ; Luo et al, 2021 , 2023 ), and have demonstrated utility in the analysis and representation of single-cell atlases (Zheng et al, 2021b ; Han et al, 2023 ; Imaz-Rosshandler et al, 2024 ). Moreover, metacells have been used to scale existing computational methods to larger datasets (Wagner et al, 2021 ; Qian et al, 2022 ; Hristov et al, 2022 ) and to enhance their robustness and interpretability (Liu et al, 2021b ; Nguyen et al, 2020 ; Magaletta et al, 2022 ; Wang et al, 2023 ; Persad et al, 2023 ; Argelaguet et al, 2022 ; Luo et al, 2021 , 2023 ).…”

“…For this reason, an aggregation of profiles of highly similar cells for GRN analysis was used even before the introduction of the metacell concept (Pliner et al, 2018 ). Currently, the metacell concept is used to enhance signal for GRN construction for scRNA-seq (Liu et al, 2021b ; Nguyen et al, 2020 ; Magaletta et al, 2022 ; Wang et al, 2023 ) and for sparser single-cell lncRNA-seq data (Luo et al, 2021 , 2023 ).…”

Building and analyzing metacells in single-cell genomics data

“…For this reason, an aggregation of profiles of highly similar cells for GRN analysis was used even before the introduction of the metacell concept 99 . Currently, the metacell concept is used to enhance signal for GRN construction for scRNA-seq 100–103 , for more sparse single-cell lncRNA-seq data 104,105 and for multi-omics data 106 . Given the computational cost of GRN analysis 17 , metacells facilitate scaling of GRN construction tools, particularly those initially developed for bulk RNA-seq 48 , to handle larger datasets 100 .…”