LIANA+_Integrating_MultiOmics_Data v1

A Protocol describing the application of LIANA+ on a spatially-resolved metabolite-transcriptome dataset from a recent murine Parkinson’s disease model Vicari et al., 2023. We demonstrate LIANA+’s utility in harmonizing spatially-resolved transcriptomics and MALDI-MSI data to unravel metabolite-mediated interactions and the molecular mechanisms of dopamine regulation in the striatum. Two particular challenges with this data are: - The unaligned spatial locations of the two omics technologies and the untargeted nature of the MALDI-MSI data, which results in a large number of features with unknown identities. Only few of which were previously identified as specific metabolites. Here, we show untargeted modelling of known and unknown metabolite peaks and their spatial relationships with transcriptomics data. Specifically, we use a multi-view modelling strategy MISTy to decipher global spatial relationships of metabolite peaks with cell types and brain-specific receptors. Then, we use LIANA+’s local metrics to pinpoint the subregions of interaction. We also show strategies to enable spatial multi-omics analysis from diverse omics technologies with unaligned locations and observations. This protocol is associated with our Nature Cell Biology paper describing LIANA+.