Computational solutions for spatial transcriptomics

Abstract

Transcriptome level expression data connected to the spatial organization of the cells and molecules would allow a comprehensive understanding of how gene expression is connected to the structure and function in the biological systems. The spatial transcriptomics platforms may soon provide such information. However, the current platforms still lack spatial resolution, capture only a fraction of the transcriptome heterogeneity, or lack the throughput for large scale studies. The strengths and weaknesses in current ST platforms and computational solutions need to be taken into account when planning spatial transcriptomics studies. The basis of the computational ST analysis is the solutions developed for single-cell RNA-sequencing data, with advancements taking into account the spatial connectedness of the transcriptomes. The scRNA-seq tools are modified for spatial transcriptomics or new solutions like deep learning-based joint analysis of expression, spatial, and image data are developed to extract biological information in the spatially resolved transcriptomes. The computational ST analysis can reveal remarkable biological insights into spatial patterns of gene expression, cell signaling, and cell type variations in connection with cell type-specific signaling and organization in complex tissues. This review covers the topics that help choosing the platform and computational solutions for spatial transcriptomics research. We focus on the currently available ST methods and platforms and their strengths and limitations. Of the computational solutions, we provide an overview of the analysis steps and tools used in the ST data analysis. The compatibility with the data types and the tools provided by the current ST analysis frameworks are summarized.

Keywords: AOI, area of illumination; BICCN, Brain Initiative Cell Census Network; BOLORAMIS, barcoded oligonucleotides ligated on RNA amplified for multiplexed and parallel in situ analyses; Baysor, Bayesian Segmentation of Spatial Transcriptomics Data; BinSpect, Binary Spatial Extraction; CCC, cell–cell communication; CCI, cell–cell interactions; CNV, copy-number variation; Computational biology; DSP, digital spatial profiling; DbiT-Seq, Deterministic Barcoding in Tissue for spatial omics sequencing; FA, factor analysis; FFPE, formalin-fixed, paraffin-embedded; FISH, fluorescence in situ hybridization; FISSEQ, fluorescence in situ sequencing of RNA; FOV, Field of view; GRNs, gene regulation networks; GSEA, gene set enrichment analysis; GSVA, gene set variation analysis; HDST, high definition spatial transcriptomics; HMRF, hidden Markov random field; ICG, interaction changed genes; ISH, in situ hybridization; ISS, in situ sequencing; JSTA, Joint cell segmentation and cell type annotation; KNN, k-nearest neighbor; LCM, Laser Capture Microdissection; LCM-seq, laser capture microdissection coupled with RNA sequencing; LOH, loss of heterozygosity analysis; MC, Molecular Cartography; MERFISH, multiplexed error-robust FISH; NMF (NNMF), Non-negative matrix factorization; PCA, Principal Component Analysis; PIXEL-seq, Polony (or DNA cluster)-indexed library-sequencing; PL-lig, padlock ligation; QC, quality control; RNAseq, RNA sequencing; ROI, region of interest; SCENIC, Single-Cell rEgulatory Network Inference and Clustering; SME, Spatial Morphological gene Expression normalization; SPATA, SPAtial Transcriptomic Analysis; ST Pipeline, Spatial Transcriptomics Pipeline; ST, Spatial transcriptomics; STARmap, spatially-resolved transcript amplicon readout mapping; Single-cell analysis; Spatial data analysis frameworks; Spatial deconvolution; Spatial transcriptomics; TIVA, Transcriptome in Vivo Analysis; TMA, tissue microarray; TME, tumor micro environment; UMAP, Uniform Manifold Approximation and Projection for Dimension Reduction; UMI, unique molecular identifier; ZipSeq, zipcoded sequencing.; scRNA-seq, single-cell RNA sequencing; scvi-tools, single-cell variational inference tools; seqFISH, sequential fluorescence in situ hybridization; sequ-smFISH, sequential single-molecule fluorescent in situ hybridization; smFISH, single molecule FISH; t-SNE, t-distributed stochastic neighbor embedding.

Graphical abstract

Fig. 1

Spatial transcriptomics methods produce data at different spatial resolutions. A)In situ ST methods detect selected targets at single-molecule resolution in their original location. Spatial information at molecular complex level localization is available. B) High-resolution spatial barcode arrays capture transcripts at subcellular resolution allowing cell organelle level localization. C) Lower resolution barcode arrays cover the area of more than one cell. Cell type analysis requires spot deconvolution methods. D) Regional illumination and collection methods offer flexibility in the target selection. The selected area can be any shape based on marker thresholding or the use of regular shapes (red or violet outline).

Fig. 2

Preprocessing of raw in situ image data. A) Image alignment is required since the corresponding signal spots in raw images from sequential probing and imaging cycles (img1, img2, img3) are not in the register in shared Euclidean space. To align the spots, the images are moved and rotated in relation to each other. B) In the aligned sequential data, the corresponding signal spots are in the register in the whole image stack, and they form the sequ-FISH barcode. C) Cell segmentation assigns every location in the image to defined cells, nuclei, or background. Transcripts are assigned to cells based on their spatial coordinates in relation to cell mask coordinates. Cells are also assigned with spatial coordinates (X, Y) in the same Euclidean space. D) Connected strings of signal spots, which are called from the image stack in panel B, are the barcodes to identify the transcript/gene at that particular coordinate location (left). The gene identities are decoded from the barcodes and counted into the cell with an overlapping coordinate location in the cell to gene matrix (right).

Fig. 3

Overview of the information that may be extracted from spatial transcriptomics data. A) Overall cell type distribution in tissue space shows specialized cell type patterns that reflect communal functions and regulation. B) Cell type-based spatial patterns and C) domains can be detected based on transcriptome clustering. D) Cellular relationships are often represented with spatial graphs of cells. The gray circles (nodes) represent the cells, while the black lines (edges) correspond to the distances between the cells. E) Cell neighborhood analysis identifies spatially connected cell type pairs in spatial domains. F) Cell–cell communication and interactions happen at many length scales, illustrated by the black arrows. G) Cell–cell communication and interactions between receptors and ligands are often detected using curated ligand-receptor lists. The green-colored cell is releasing ligands (blue dots), while the green cell is communicating with the yellow-colored cell directly by receptors. H) Subcellular distribution of transcripts (orange dots) can be used in ST analysis, to understand the cellular scope.