MAPLE: A Hybrid Framework for Multi-Sample Spatial Transcriptomics Data

Abstract

High throughput spatial transcriptomics (HST) technologies provide unprecedented opportunity to identify spatially resolved cell sub-populations in tissue samples. However, existing methods preclude joint analysis of multiple HST samples, do not allow for differential abundance analysis (DAA), and ignore uncertainty quantification. To address this, we developed MAPLE: a hybrid deep learning and Bayesian modeling framework for joint detection of spatially informed sub-populations, DAA, and uncertainty quantification. We demonstrate the capability of MAPLE to achieve these multi-sample analyses through four case studies that span a variety of organs in both humans and animal models. An R package maple is available on GitHub at https://github.com/carter-allen/maple .