Estelle Duguet
Pronouns: she/her
Research Mentor(s): Matthew Patrick
Research Mentor School/College/Department: Department of Dermatology / Medicine
Program: UROPF
Session: Session 4 (1:40pm – 2:30pm)
Authors: Lam C Tsoi, Haihan Zhang, Johann E Gudjonsson
Abstract
Generalizability, portability, and robustness are key in analyzing high dimensional data, and specific data types require unique data management. Therefore, having a single, generalized pipeline will provide efficient analysis and facilitate researchers to provide interpretation and biological inference for their data. High resolution spatial expression profiling is an advanced technology, and allows us to image and investigate tissues at subcellular resolutions. Using CosMX and data on different skin conditions, we developed a pipeline that analyzes, performs cell segmentation and groups these cells based on genetic compositions or features. Our implementation enables users to provide a config file and supports clear instructions that ensure reproducibility. This makes it so that a single script can be used on multiple data types. Generalizing these scripts into a software library unifies data presentations and benefits other researchers in the field. In the future, the library will be open source on GitHub, for easy access for all researchers.
