Abstract 2335: MultiNMF: multiview factorization for joint modeling of spatial multi-omics and histology images

Abstract

Abstract An increasing number of cancer research studies employ spatially resolved transcriptomics (SRT) to investigate the composition of tumor microenvironment in a cancer type of interest. These studies have defined tumor microenvironment (TME) states and spatial domains based on clustering spatial gene expression patterns in SRT in an unbiased manner, yet a more thorough delineation of TME states requires the incorporation of the tumor’s histology image. Here, we develop MultiNMF, a multiview factorization approach that is suitable for cancer research studies where joint profiles of spatial multi-omics and tumor histology images are available. We apply MultiNMF to analyze a set of TNBC SRT primary tumor samples and reveal TME states in the stromal, epithelial, and immune enriched compartments, defined by distinct histomorphological features. We further illustrate the ability of the approach to extend to paired spatial ATAC-seq and histology dataset that is recently published on HER2 breast cancer. MultiNMF thus permits an automated and data-driven decomposition of SRT and spatial ATAC data supported by histomorphological evidence. Context In SRT by 10X Visium, the hematoxylin eosin (H&E) staining image is automatically aligned to the slide where the tissue section is mounted. Image patches can be easily extracted from areas under the SRT barcoded spots. For these images, I will apply a pre-trained convolutional neural network (CNN) model to extract high dimensional features from spot images. The image features can be added as a second view after gene expression view for joint analysis under MultiNMF. Results MultiNMF can factorize the SRT data into components well-supported by histological evidence. It has identified T-cell infiltrating regions, and EMT-enriched regions with distinct immune and stroma morphological characteristics. The T-cell infiltration neighbors a region that has an appearance of necrosis according to the pathologist evaluation. These domains further demonstrate enrichments of modules with motif enrichment (known as regulons) according to SCENIC analysis. Analysis of spatial ATAC illustrates not only HER2 but also its amplicon amplification. From MultiNMF components of spatial ATAC, we derive regulatory regions for T-cell/B-cell marker genes. Overall, MultiNMF is a novel model that has not been applied to SRT field. It provides an multiomic extension to the traditional NMF. Citation Format: William Bowie, Stacy Wang, Benjamin Strope, Qian Zhu. MultiNMF: multiview factorization for joint modeling of spatial multi-omics and histology images [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2335.