TIST: Transcriptome and Histopathological Image Integrative Analysis for Spatial Transcriptomics

Sequencing-based spatial transcriptomics (ST) is an emerging technique to study in situ gene expression patterns at the whole-genome scale. In addition to transcriptomic data, the technique usually generates matched histopathological images for the same tissue sample. ST data analysis is complicated by severe technical noise; matched histopathological images with high spatial continuity and resolution introduce complementary cellular phenotypical information and provide a chance to mitigate the noise in ST data. Hence, we propose a novel ST data analysis method called transcriptome and histopathological image integrative analysis for spatial transcriptomics (TIST), which integrates the information from sequencing-based ST data and histopathological images. TIST uses a Markov random field (MRF) model to learn the macroscopic cellular features from histopathological images and devises a random-walk-based strategy to integrate the extracted image features, the transcriptomic features and the location information for spatial cluster (SC) identification and gene expression enhancement. We tested TIST both on simulated datasets and on 33 real datasets; we found that TIST achieved superior performance on multiple tasks, which illustrates the utility of this method in facilitating the discovery of biological insights from sequencing-based ST data.