Practical considerations for machine learning-enabled discoveries in spatial transcriptomics

Development and homeostasis in multicellular systems require exquisite control over spatial molecular pattern formation and maintenance. Advances in spatially-resolved and high-throughput molecular imaging methods such as multiplexed immunofluorescence and spatial transcriptomics (ST) provide exciting new opportunities to augment our fundamental understanding of these processes in health and disease. The large and complex datasets from these techniques, particularly ST, have led to the rapid development of innovative machine learning (ML) tools in this domain. These tools are now increasingly featured in integrated experimental and computational workflows to disentangle signals from noise in complex biological systems. Methods development for spatial biology data analysis draws on diverse subfields and perspectives, and it can be challenging to understand the primary computational considerations of ML in ST. To address this, we highlight some ST analysis goals that ML can help address. We also describe four major data science concepts and related heuristics that can help practitioners choose the right tools for the right biological questions.