M-Band Wavelet-Based Imputation of scRNA-seq Matrix and Multi-view Clustering of Cell

Abstract

Wavelet analysis has been recognized as a cutting-edge and promising tool in the fields of signal processing and data analysis. However, application of wavelet-based method in single-cell RNA sequencing (scRNA-seq) data is little known. Here, we present M-band wavelet-based imputation of scRNA-seq matrix and multi-view clustering of cells (WIMC). We applied integration of M-band wavelet analysis and uniform manifold approximation and projection (UMAP) to a panel of single cell sequencing datasets by breaking up the data matrix into a trend (low frequency or low resolution) component and (M-1) fluctuation (high frequency or high resolution) components. We leverage a non-parametric wavelet-based imputation algorithm of sparse data that integrates M-band wavelet transform for recovering dropout events of scRNA-seq datasets. Our method is armed with multi-view clustering of cell types, identity, and functional states, enabling missing cell types visualization and new cell types discovery. Distinct to standard scRNA-seq workflow, our wavelet-based approach is a new addition to resolve the notorious chaotic sparsity of scRNA-seq matrix and to uncover rare cell types with a fine-resolution. Author summaryWe develop M-band wavelet-based imputation of scRNA-seq matrix and multi-view clustering of cells. Our new approach integrates M-band wavelet analysis and UMAP to a panel of single cell sequencing datasets via breaking up the data matrix into a trend (low frequency or low resolution) component and (M - 1) fluctuation (high frequency or high resolution) components. Our method enables us to efficiently impute sparse scRNA-seq data matrix and to examine multi-view clustering of cell types, identity, and functional states, potentializing missing cell types recovery, fine rare cell types discovery, as well as functional cell states exploration.