Spatially structured cell heterogeneity within tissues is essential for healthy organ function. This heterogeneity arises from a tightly regulated interplay of cell proliferation, cell differentiation, and spatial organization during development. To study these diverse processes, developmental biology has become a data-intensive science through the use of high-throughput imaging and multi-omics technologies. To take advantage of this wealth of data, we have developed several inference and machine learning methods that aim to bridge the study of morphogenesis with genome-wide gene expression patterns. We will illustrate our approaches using several model systems including the mouse, C. elegans development, and Drosophila oocytes. In particular, we will focus on Spatial Principal Component Analysis as a relevant method to analyse multivariate spatial patterns in epithelial tissues and to compare different conditions, at different time points and with different genetic perturbations. Taken together, our results show that integrating insights from cell-scale feature patterning and mechanical stress provides new insights into morphogenesis.