Coupled Co-clustering-based Unsupervised Transfer Learning for the Integrative Analysis of Single-Cell Genomic Data

Abstract

Unsupervised methods, such as clustering methods, are essential to the analysis of singlecell genomic data. Most current clustering methods are designed for one data type only, such as scRNA-seq, scATAC-seq or sc-methylation data alone, and a few are developed for the integrative analysis of multiple data types. Integrative analysis of multimodal singlecell genomic data sets leverages the power in multiple data sets and can deepen the biological insight. We propose a coupled co-clustering-based unsupervised transfer learning algorithm (coupleCoC) for the integrative analysis of multimodal single-cell data. Our proposed coupleCoC builds upon the information theoretic co-clustering framework. We applied coupleCoC for the integrative analysis of scATAC-seq and scRNA-seq data, sc-methylation and scRNA-seq data, and scRNA-seq data from mouse and human. We demonstrate that coupleCoC improves the overall clustering performance and matches the cell subpopulations across multimodal single-cell genomic data sets. The software and data sets are available at https://github.com/cuhklinlab/coupleCoC

coupleCoC+: an information-theoretic co-clustering-based transfer learning framework for the integrative analysis of single-cell genomic data

“…n T and n S are the numbers of cells in the target data and the source data, correspondingly. Because we included the unlinked features when implementing CoC, k-means, Cusanovich2018, cisTopic, SC3, SIMLR and BPRMeth-G, the clustering results for these methods are better than that presented in Zeng and Lin (2020). Example 2: integrative clustering for mouse and human scRNA-seq data…”

“…(5) As mentioned before, the two terms T (C Y , C Z ) and S (C X , C Z ) in formula ( 5) share the same feature cluster C Z , which can be viewed as a bridge to transfer knowledge between the source data and the target data (Dai et al, 2008;Zeng and Lin, 2020). The dimension of the feature space shared by the source data S and the data T is reduced by clustering and aggregating similar features.…”

“…The intuition for choosing N sub is that the KL divergence D KL (p T (Ỹ h T,N sub ,Z T )||p S (X h S,N sub ,Z S )) will be larger if the clusters being matched are less similar when they are forced to be matched. N sub is chosen similarly as in Zeng and Lin (2020). More details are given in Section 1.3 in Supplementary Materials.…”

“…In recent years, some methods are developed for this purpose. They include Seurat (Butler et al, 2018;Stuart et al, 2019), MOFA (Argelaguet et al, 2018), coupleNMF (Duren et al, 2018), scVDMC (Zhang et al, 2018), LIGER (Welch et al, 2019), scACE (Lin et al, 2019), MOFA+ (Argelaguet et al, 2020), scAI (Jin et al, 2020), coupleCoC (Zeng and Lin, 2020) and scMC (Zhang and Nie, 2021). A more comprehensive discussion on integration of single-cell genomic data is presented in David et al (2020).…”

“…coupleCoC+ also performs matching of a subset of cell clusters across the source data and the target data, which may represent shared cell types across the two datasets. We refer our model as coupleCoC+, because it is based on the framework of our previously proposed coupleCoC (Zeng and Lin, 2020). coupleCoC+ addresses the limitations of coupleCoC by including the unlinked features in target data and it also integrates co-clustering and cell type matching in one step for better use of information from the source data.…”