Community Highlight: Grace Feng
My name is Grace Feng (GF) and I’m an upcoming junior at the University of Michigan studying computer science.
Grace Feng, University of Michigan
I joined the KBase team during the Summer 2025 term as part of the Research Student Internships program at ORNL. I developed the Normalize Phenotypes for GWAS app to be used in the preprocessing step of the Genome Wide Association Studies (GWAS) pipeline. The app accepts continuous phenotypic data and has two modes: view and edit. In view mode, the user can view the distributions of each phenotype, to get a better understanding of the necessary transformation or outlier removal required. In the editing stage, the user can run the app again,selecting the phenotypes they would like to edit. The app then returns the phenotypes that were selected and changed, which the user can use in downstream analysis. I was able to show, through applying the app’s Box-Cox transformation to sample data of Populus trichocarpa (Black Cottonwood), the significant changes in the GWAS test result.
How has using KBase supported your research or project?
(GF) KBase has provided me with the platform and resources to make this project happen. The Software Development Kit (SDK) was well documented and allowed me to seamlessly integrate the Python packages I needed to use into a KBase module.
How have you benefited from working with KBase?
(GF) As a student starting out in my computer science education, working with KBase and the SDK has been so beneficial for me as it was a chance for me to build a project from the ground up, tailor it to what the group needed, and update the app iteratively. It was the perfect step up from my class projects, which had a lot more rigid instructions. Working on the normalization app gave me flexibility and allowed me to gain so much practical experience coding in Python.
What is your favorite part about using KBase?
(GF) I like the structure of the Narrative interface. It reminds me of Jupyter Notebooks – I like how you can toggle between the input and the visualizations.
What is one “behind the scenes” thing you want to share about your project – e.g., from collecting the data, running the analysis, publishing the results? Or, any tips on using the tool, learning more about your project, or ways to get involved?
(GF) One thing that I struggled with behind the scenes was performing the transformations on the AttributeMapping type. Because attribute mappings store results row by row and I needed to perform calculations on columns, there was some data manipulation required.
What role do you see KBase having in open science principles (related to your work)?
(GF) During my time as an undergraduate researcher at Michigan, I performed Mendelian Randomization, which is a post-GWAS analysis. I remember wishing there was an easier way to share my pipeline and display the results. I see KBase being a very powerful tool to do this, because anyone with a free account can view public narratives and see exactly what data and tests you are running.
Interested in the RSI program at ORNL? Apply here: https://education.ornl.gov/rsi/
Follow Grace on LinkedIn: https://www.linkedin.com/in/grace-feng24
