Abstracts Communications and the Arts

SpotComm: um novo algoritmo para caracterização da comunicação celular em transcriptômica espacial

by Marcelo Luiz Brunatto Falchetti

Abstract: Since cell communication, at least in an autocrine and paracrine fashion is spatially limited, spatial transcriptomics methods can increase the reliability of predicted communications. Cell communication analysis authors have been thinking about this and developing algorithms that take advantage of spatial information in their predictions. However, algorithms for analyzing cell communication in spatial transcriptomics data have so far focused on the analysis of spot populations, the cluster-based analyses, whether transcriptional clusters or tissue regions, and, that said, do not examine the communication particularities of the spot transcriptome, which, depending on the technology, can be one cell or a small group of cells. Based on this, we have developed a cell communication prediction algorithm that uses coordinate data, of spots, as the unit of analysis, the SpotComm. This algorithm has functions capable of defining intercellular, intracellular communications and intracellular signaling based on the presence and correlation of transcripts in spatial transcriptomics data and is able to integrate spatial transcriptomics data with paired or unpaired scRNA-seq data. SpotComm is capable of handling spatial transcriptomics data in single-section (2D) or multi-section (3D) analyses, provide analyses of entire transcriptional clusters and/or contact zones between clusters, and provide scenarios of communication and signaling between monomeric interacting proteins and homomultimeric and heteromultimeric protein complexes. The algorithm is able to generate data and provide metadata to the user such as the presence, proportion and expression of interactors and the reference and curation of communication and signaling. It also provides the proportion of cells and proportion of co-occurrence, by cell type, with detection of the elements of intracellular signaling pathways. Using SpotComm we were able to predict communications that are known in tertiary lymphoid structures and in tumor areas profiled by type I interferon responses in breast cancer and predicted the cellular elements viable for signaling. Furthermore, we detected several potential inter- and intracellular communications not detected in cluster-based analysis that may be important in understanding homeostasis but also in disease prognosis, diagnosis and treatment.