Dictionary learning for integrative, multimodal, and scalable single-cell analysis

Abstract

Mapping single-cell sequencing profiles to comprehensive reference datasets represents a powerful alternative to unsupervised analysis. Reference datasets, however, are predominantly constructed from single-cell RNA-seq data, and cannot be used to annotate datasets that do not measure gene expression. Here we introduce 'bridge integration', a method to harmonize single-cell datasets across modalities by leveraging a multi-omic dataset as a molecular bridge. Each cell in the multi-omic dataset comprises an element in a 'dictionary', which can be used to reconstruct unimodal datasets and transform them into a shared space. We demonstrate that our procedure can accurately harmonize transcriptomic data with independent single cell measurements of chromatin accessibility, histone modifications, DNA methylation, and protein levels. Moreover, we demonstrate how dictionary learning can be combined with sketching techniques to substantially improve computational scalability, and harmonize 8.6 million human immune cell profiles from sequencing and mass cytometry experiments. Our approach aims to broaden the utility of single-cell reference datasets and facilitate comparisons across diverse molecular modalities. Availability: Installation instructions, documentations, and vignettes are available at http://www.satijalab.org/seurat

Depth normalization for single-cell genomics count data

“…In contrast, sparse matrices have been used for high-performance computing for a long time (Orchard-Eays 1956; Markowitz 1957), and can drastically reduce the memory overhead required to perform memory-intensive computations. While recently developed “sketching” procedures (Hao et al 2022) that subsample matrix operations for scalable computation may provide workarounds for dense matrices, we believe that sparsity will remain an important consideration for normalization transformations for the foreseeable future.…”

Integrated Single-Cell Genotyping and Chromatin Accessibility ChartsJAK2V617FHuman Hematopoietic Differentiation

“…3h ). Cluster assignments were further supported via orthogonal annotation utilizing cell mapping through a novel multiomic bridge integration approach 54 ( Extended Data Fig. 3i; see materials and methods ).…”

“…h , Gene accessibility scores for key stem cell ( HLF, AVP, CD34, CD38 ) marker genes across clusters HSPC1 and HSPC2 (Wilcoxon rank sum test), consistent with HSPC1 representing earlier HSPCs (higher HLF, AVP and CD34 ) and HSPC2 representing later HSPCs (higher CD38 and lower HLF, AVP and CD34 ). i , Confusion matrix between manually annotated cluster labels and predicted labels based on scRNA-seq reference via bridge integration mapping 54 (see materials and methods).…”

Integrated single-cell (phospho-)protein and RNA detection uncovers phenotypic characteristics of human antibody secreting cells

“…Computational strategies encounter several considerations as how to define anchors, scalability and handling missing data ( 43 ). Several of these challenges are being addressed by recently developed tools including MOFA+ ( 24 ), multiVI ( 44 ), COBOLT ( 45 ), StabMap ( 46 ) scMVP ( 47 ), and Bridge Integration ( 48 ). So-far there was no illustration of integrating mRNA datasets with a comprehensive intracellular phospho-protein and transcription factor dataset using a common set of surface proteins.…”