Here we introduce ‘bridge integration', which performs integration of single-cell datasets measuring different modalities by leveraging a separate dataset where both modalities are simultaneously measured as a molecular ‘bridge’. The multi-omic bridge dataset, which can be generated by a diverse set of technologies23-32 (Fig. 1a), helps to translate information between disparate measurements, resulting in robust integration without requiring any limiting biological assumptions. We illustrate the broad applicability of our approach, demonstrating its performance across five different molecular modalities, and highlighting specific requirements for the multi-omic dataset that can help to guide experimental design.
