Skip to main content
Fig. 1 | Genome Medicine

Fig. 1

From: Machine learning integrative approaches to advance computational immunology

Fig. 1

A Multimodal immunological datasets can comprise multiple assays across different modalities and resolutions. The number of features measured in each assay ranges from tens to hundreds of thousands. Some of these assays (CITEseq [59]; Multiome [60]) collect joint information from the same cells or samples (modalities aligned vertically). Different assays may share subsets of features (CyTOF [13], FACS [12]) (modalities aligned horizontally). Sample level measurements do not have a cellular resolution (ATAC-seq [61], RNA-seq [62], mass spectrometry [63]; BCR and TCR sequencing [64]) but can be performed in parallel to single-cell assays; spatially resolved cells can be extracted from multiple platforms, sequencing or imaging-based (Spatial ATAC [38], Spatial CITESeq [65], Visium [66], MERFISH [67]). B Spatial profiling carries information about RNA expression at individual spatial barcodes (BC). Single-cell references can be leveraged to deconvolute spatial data inferring cell type proportions and gene expression at spatial locations. Histological sections are often an accompanying assay. They can be segmented to recover cell and subcellular structures, as well as general tissue properties such as the morphology of cells, the density of cells at specific locations and cell-to-cell interactions

Back to article page