Fatemah Alsolaiman

تخصص المعلوماتية الحيوية · دفعة 2024

Evaluation of Transcript-Guided Cell Segmentation on GBM-Derived Single Molecule Spatial Ranscriptomics Data

Glioblastoma (GBM) is a heterogeneous and aggressive brain tumor characterized by a variety of genomic mutations and aberrant gene expression patterns. This molecular heterogeneity significantly influences tumor development and progression. However, tumor progression is also a result of cells forming communities due to their spatial relationships, which gives rise to functional and behavioral attributes to promote survival and resistance. Thus, understanding spatial properties is critical to developing targeted therapies and improving patient outcomes. Accurate single-cell segmentation is vital to measure the spatial distribution of GBM gene expression and tumor cell organization with their surrounding microenvironment and at multiple spatial scales. To address this, we are implementing, testing, and evaluating different segmentation methods to determine the best performing algorithms. This project will compare two different segmentation methods with original Xenium segmentation masks- Baysor and Proseg. This comparison involves integrating segmentation masks with images, annotating randomly selected regions, and evaluating their accuracy and suitability. The expected outcome of this project is a comprehensive evaluation of segmentation tools, resulting in improved methods for analyzing spatial gene expression data in GBM. By advancing our understanding of tumor architecture, this study offers an opportunity to improve the prognosis and treatment outcomes of patients with GBM.

اقرأ التقرير النهائي (PDF)

اطّلع عليه في تقرير البرنامج (PDF)