- Genomic Data-Guided Computational Modeling of Cancer
- Project Year:
2023
- REU Student (s):
Elm Markert | Smith College MA
- Student 1 Institution:
Smith College
- Project Mentor:
Subhajyoti De
- Project Mentor Area:
Rutgers Cancer Institute of New Jersey
- Project Abstract:
Tumors grow from somatic cells inside the human body, often undetected without any major symptoms. But tumors shed cell-free tumor DNA (cfDNA), as well as proteins, intact tumor cells, and other molecules into the blood and other bodily fluids. As cfDNA travels through the bloodstream, it is degraded by various nucleases, salt, etc. such that when blood is drawn from patients for liquid biopsy for diagnostic purposes, the cfDNA has been fragmented into smaller pieces. The methylation, length, and molecular signatures of these cfDNA fragments have the ability to provide information about the location and type of cancer. However, the methods through which different types of cfDNA are degraded remain a mystery, and there is currently no broad, all-encompassing cancer diagnostic tool that utilizes this data. We utilize a mathematical technique called non-negative matrix factorization (NMF) to identify predominant characteristics of end-sequences of cfDNA from cancer patients with multiple types of cancer.
