Sharifa S
- Research Program Mentor
PhD candidate at Harvard University
Expertise
Python, R, C++, Bioinformatics, Computational Biology, Systems Biology, Machine Learning, Computer Vision, Data Science, Fellowship applications, College applications
Bio
I am a current PhD student at Harvard University in the Systems, Synthetic and Quantitative Biology program. I have a dual bachelor’s degree in computer science and biology from Union College. Prior to attending Harvard I was a software engineer at Google for one year in their Software Engineering Residency program. I have held competitive summer internships at the NASA Jet Propulsion Laboratory and at Stanford Medical School. I was also awarded the Klemm Fellowship to research sharks underwater in Fiji. I have experience with prestigious project based competitions including winning a Major League Hacking Hackathon and being a finalist for the ArtScience Prize.Project ideas
Evolutionary genomic study of shark cartilage mineralization
There has been a recent flurry of whole genome sequencing of sharks such as the great white shark (2019), the brownbanded bamboo shark (2018), cloudy catshark (2018) and whale shark (2018). The new availability of these genomes makes it a prime candidate for study. Sharks are fascinating and slow evolving creatures. You can choose to focus on genes related to cartilage and bone development, aligning the shark whole genomes to known cartilage forming genes and bone forming genes. Cartilaginous fish such as sharks and bony fish diverged over 450 million years ago. Cartilage is the precursor to bone. Comparing these two groups could create better insight into the evolution of bone mineralization genes. For a more translational project, you may choose to align the shark genomes to the mouse or human genomes and look into genes involved in cartilage or bone development diseases.
Machine learning based genomic and pathology image analysis on the Cancer Genome Atlas
The Cancer Genome Atlas (TCGA) is a wealth of open source data including patient health records, genomic sequencing and histology slides. Focusing on a rare cancer would be ideal for this project as they tend to be understudied and even analyses utilizing small datasets could lead to interesting discoveries. Utilizing machine learning techniques we can analyze this data to predict correlations between morphological histology features and mutations, patient survival based on histology or genomic data, etc. My research at the Pathology Image Analysis lab at the Brigham Women’s Hospital is focused on creating and applying deep learning methods to pathology image analysis. There are multiple open sources tools developed by the lab such as CLAM that could be utilized for this project. Website for TCGA: https://portal.gdc.cancer.gov/ Website for CLAM: https://github.com/mahmoodlab/CLAM