Abstract 7430: Development and demonstration of a user-friendly application to explore single-cell atlas for prostate cancer (PCa) and analyze newly generated single-cell and spatial PCa datasets
Shivang Sharma,Eugene Shenderov
DOI: https://doi.org/10.1158/1538-7445.am2024-7430
IF: 11.2
2024-03-28
Cancer Research
Abstract:Background: Single-cell atlases provide a high-resolution and tissue-specific depiction of cellular state at the transcriptomic level. Newer technologies like spatial transcriptomics, MALDI, and long-read sequencing introduce orthogonal modalities to single-cell data, making the analyses of large atlases even more complex. RShiny package in R is used to develop user-friendly GUI applications that could enable users to use analysis tools in an interactive manner, thus minimizing the requirement of prior programming experience. Methods: Publicly available scRNA-seq datasets were collected for a wide range of prostate-related pathologies. The datasets were evaluated for their quality based on cell count and availability of the appropriate markers in each compartment. Finally, 11 datasets out of 16 datasets were selected for building integrated datasets. These datasets were filtered for cell types with <10% mitochondrial RNA, and cells with <500 genes were removed from the datasets. The datasets were then clustered to evaluate the effects of cell cycle, dissociation-induced stress, and the presence of doublets. The quality-controlled datasets were batch-corrected using the Harmony package and the harmonized principal components were used in Seurat for clustering the integrated data. Tools used for the analysis of single-cell datasets, beginning from raw counts, have been developed in R with a focus on modularity. These tools are packaged into a simple and intuitive RShiny application which follows the workflow shown in Figure 1a. Results: The atlas obtained include samples from healthy prostate, benign prostate hyperplasia, primary prostate tumor, ICC/IDC, castration-resistant prostate cancer (CRPC), and Neuroendocrine prostate cancer (NEPC). The dataset includes 72 patients and ~320k cells with a median number of ~1600 genes expressed. We were able to distinctly identify Luminal, Hillock, Club, Basal, Neuroendocrine, Endothelial, Fibroblast, Smooth Muscle, Myeloid, T, B, NK, Mast, and Plasma cells in the integrated along with some unknown cellular subtypes in the epithelial compartment. The integrated data was manually annotated using cell type-specific markers. The single-cell atlas obtained was weaved with the workflow tools into an RShiny app. Conclusions: Besides being a useful reference for PCa, the atlas would find extensive use in understanding the transcriptomic landscape of PCa biology. The integrated dataset will also prove crucial in developing bulk RNA sequencing deconvolution and single-cell imputation algorithms. Finally, we aim to provide a user-friendly interface to the prostate cancer community for using the single-cell PCa atlas and analysis tools with minimal computational requirements. Citation Format: Shivang Sharma, Eugene Shenderov. Development and demonstration of a user-friendly application to explore single-cell atlas for prostate cancer (PCa) and analyze newly generated single-cell and spatial PCa datasets [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular s); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl) nr 7430.
oncology