LSMMD-MA: scaling multimodal data integration for single-cell genomics data analysis

Abstract

Motivation: Modality matching in single-cell omics data analysis-i.e. matching cells across datasets collected using different types of genomic assays-has become an important problem, because unifying perspectives across different technologies holds the promise of yielding biological and clinical discoveries. However, single-cell dataset sizes can now reach hundreds of thousands to millions of cells, which remain out of reach for most multimodal computational methods.

Results: We propose LSMMD-MA, a large-scale Python implementation of the MMD-MA method for multimodal data integration. In LSMMD-MA, we reformulate the MMD-MA optimization problem using linear algebra and solve it with KeOps, a CUDA framework for symbolic matrix computation in Python. We show that LSMMD-MA scales to a million cells in each modality, two orders of magnitude greater than existing implementations.

Availability and implementation: LSMMD-MA is freely available at https://github.com/google-research/large_scale_mmdma and archived at https://doi.org/10.5281/zenodo.8076311.

Figure 1.

(A) Fastest MMD-MA variant as a function of the number of samples and the number of features. The black region in the top right corner means that all variants ran out of memory. (B) Runtime as a function of number of cells for different implementations of MMD-MA, when the dimension p of the input data varies. The black and green dotted lines with cross markers correspond to the original implementations of MMD-MA as written by  Liu et al. (2019) (black) and Singh et al. (2020) (green). The runtime for different values of p, from 100 to 10 000, is shown in Supplementary Appendix Fig. A1.