Powerful and accurate case-control analysis of spatial molecular data

As spatial molecular data grow in scope, there is a pressing need to identify disease-associated spatial structures. Current approaches typically make restrictive assumptions such as representing tissue regions by abundances of discrete cell types and samples by abundances of discrete niches; this risks overlooking important signals. I will discuss variational inference-based microniche analysis (VIMA), a method combining deep learning with principled statistics to discover disease-associated spatial features with greater flexibility and precision. VIMA trains an ensemble of variational autoencoders to summarize the contents of every small tissue patch in a dataset via numeric “fingerprints”. It uses these to define many data-dependent, overlapping “microniches” and meta-analyzes them to identify microniches whose abundance correlates significantly with case-control status. After describing the method, I’ll show how it performs in simulations designed to assess calibration, power, and spatial accuracy. I’ll then give some examples of how we have used VIMA on spatial datasets from diverse spatial modalities, where it recapitulates known biology and identifies novel spatial features of disease.