AbstractSingle-cell multi-omics sequencing data can provide a comprehensive molecular view of cells. However, effective approaches for the integrative analysis of such data are challenging. Although achieved state-of-the-art performance on single-cell multi-omics data integration and did not require any correspondence information, either among cells or among features, current manifold alignment based integrative methods are often limited by requiring that single-cell datasets be derived from the same underlying cellular structure. To overcome this limitation, we present Pamona, an algorithm that integrates heterogeneous single-cell multi-omics datasets with the aim of delineating and representing the shared and dataset-specific cellular structures. We formulate this task as a partial manifold alignment problem and develop a partial Gromov-Wasserstein optimal transport framework to solve it. Pamona identifies both shared and dataset-specific cells based on the computed probabilistic couplings of cells across datasets, and it aligns cellular modalities in a common low-dimensional space, while simultaneously preserving both shared and dataset-specific structures. Our framework can easily incorporate prior information, such as cell type annotations or cell-cell correspondence, to further improve alignment quality. Simulation studies and applications to four real data sets demonstrate that Pamona can accurately identify shared and dataset-specific cells, as well as faithfully recover and align cellular structures of heterogeneous single-cell modalities in the common space. Pamona software is available at https://github.com/caokai1073/Pamona.
Manifold alignment for heterogeneous single-cell multi-omics data integration using Pamona
