Jeremy Philip D’Silva

While single-cell RNA sequencing has advanced our understanding of cell fate, identifying molecular hallmarks of potency-a cell's ability to differentiate into other cell types-remains a challenge. Here we introduce CytoTRACE 2, an interpretable deep learning framework for predicting absolute developmental potential from single-cell RNA sequencing data. Across diverse platforms and tissues, CytoTRACE 2 outperformed previous methods in predicting developmental hierarchies, enabling detailed mapping of single-cell differentiation landscapes and expanding insights into cell potency.

Abstract Single-cell transcriptomics enables the study of heterogeneity by uncovering the transcriptome of individual cells, which is averaged out in bulk analyses. This facilitates the grouping of cells based on similar transcriptomic profiles, potentially identifying cell subpopulations, such as rare tumor cells that mediate drug resistance or undergo metastasis. Image-based transcriptomic techniques can go a step further in uniquely profiling cells, by providing information about the spatial co-ordinates of transcripts and cells, whilst shedding light on cell morphology for phenotypic characterization. Single-molecule Fluorescence in situ hybridization (smFISH) is one image-based method in which fluorescently labeled tiling-oligonucleotide (TO) probes are bound along the length of complementary RNA strands, allowing for the detection of transcripts at nanometer-scale resolution. However, smFISH data analysis has been difficult to implement at scale due to the tradeoff between transcript resolution and imaging area, as well as the lack of automated feature detection tools. To address this issue, we have developed a pipeline called HIgh-Throughput Single-molecule FISH (HITSFISH), which uniquely combines smFISH with structured-illumination microscopy allowing for large-area high-resolution imagery with automated feature detection and analysis. We have validated our platform by showing high transcriptomic concordance between HITSFISH and other standard methods such as bulk RNA-seq and qPCR, demonstrating its ability to accurately measure RNA levels. With our pipeline we look to answer questions pertaining to the spatiotemporal landscape of gene expression within and across normal and cancerous cells. In addition, by using immunofluorescence in tandem with our platform, we can study transcript localization at various subcellular landmarks, such as membraneless organelles, to gain insight on subcellular mechanisms of gene regulation. Citation Format: Andrej Coleski, Sethu Pitchiaya, Jeremy D'Silva, Marcin Cieslik, Arul Chinnaiyan. Spatially resolved image-based transcriptomics using high-throughput single-molecule fluorescence in situ hybridization (HITSFISH) [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2162.