“…As a future development, deep learning shows exciting promise for linking multiomic and spatial data across biological scales and formats (Bersanelli et al, 2016;Haas et al, 2017;Mirza et al, 2019;Nguyen and Wang, 2020). By registering genome-wide single-cell sequencing data to sparse spatial transcriptomic reference frames, deep learning computer vision methods predict spatial expression patterns with increased coverage, error reduction, and multiomic integration (Biancalani et al, 2020;Ma et al, 2020). In principle, similar computational strategies could integrate target-specific proteomic datasets with mass spectrometry imaging (MSI; Xu and Li, 2019) or multi-round protein imaging for spatial proteomics.…”