Computation of cancer gene expression signatures in spatial transcriptomics data

Even though tumor originates from clones of cells, it develops a substantial intratumor heterogeneity in terms of cellular morphology, gene expression, proliferative and metastatic potential. Due to this heterogeneity, diagnostic and prognostic gene expression cancer signatures often fail in the evaluation of bulk gene expression tumor profiles. Moreover, the internal organization of tumors has potential consequences on treatment response and resistance. Therefore, discerning the complexity of both composition and internal structure could provide a valid step towards the understanding of tumor biology. Spatial transcriptomics is a new approach in the analysis of transcriptomes that allows the analysis of gene expression level in the intact tissue, maintaining the spatial information. During my thesis project I worked in the application of cancer gene expression signatures on spatial breast cancer transcriptome data, highlighting how the resulting panel of spatially resolved cancer gene expression scores provide powerful information in tumor data interpretation.