NCSI: 3D Visualization, Identification, and Quantification of Biological Cells using Virtual Reality (Intern Position)

Status	Completed
Mentor Name	Matthew Niepielko
Mentor's XSEDE Affiliation	Mentor
Mentor Has Been in XSEDE Community	Less than 1 year
Project Title	3D Visualization, Identification, and Quantification of Biological Cells using Virtual Reality (Intern Position)
Summary	This project will involve two undergraduate students developing a novel computational tool to advance our capabilities to visualize and quantify biological cells in 3D using virtual reality. Biological data such as images of cells developing in a tissue are stored as layered 2D images gathered on instruments such as confocal microscope; this data acquisition creates pitfalls when attempting to quickly identify and quantify 3D objects such as cells that are condensed within a finite space during development or tumorigenesis. To circumvent these issues, our goal is to create an open source tool using the Unity Visualization Toolkit (UVT) that can 1) render 2D layered imaging data as distinct 3D objects in virtual reality (VR), 2) identify individual cells within a condensed space, and 3) automatically quantify cell populations.
Job Description	The students involved in this project will be co-mentored by Dr. David Joiner (Computational Science) and Dr. Matthew Niepielko (Computational Biology). Each student will 1) become proficient in computer program using the Unity Visualization Toolkit, 2) implement algorithms for 3D object detection, 3) generate an automated analysis pipeline that quantifies biological data in 3D that is stored as layered 2D images, 4) acquire scientific communication skills through writing, presentation and teamwork, and 5) learn how computational approaches can be integrated with biological research to help solve complex biological questions through collaboration. Current progress on this project has led to a “proof of concept” 3D visualization tool and our student’s goal includes automating the analysis process and implementing new algorithms that will speed up the current analysis pipeline.
Computational Resources	This project will involve development of image analysis tools for large confocal microscope datasets, including computer vision algorithms applied to 3D dataset analysis. Our goals include a mixture of traditional machine learning approaches in 2 and 3D, as well as applying neural networks to data arranged as 2 and 3D structures. We expect that the computational resources needed to test, and train models will be aided using XSEDE resources. Moving from 2D to 3D neural networks will increase our computational complexity, as typically this increases our image data amount by 2 orders of magnitude, and our problem poses challenges both in amount and structure of data compared to traditional computer vision problems.
Contribution to Community	Completion of this project will yield novel computer software for the visualization and quantification of biological cells that form within a developing tissue. Quantification of developing cells is needed in biological studies to determine genetic links and treatments for developmental disorders, diseases such as cancer, and tumor development. Current tools available to the biological community for such precise measurements are costly, inadequate, and/or are not adaptable for specific application. Our project will provide the community with a new tool that will 1) be adaptable to many studies and data types, 2) open source and free to use, and 3) provides non-bias and accurate cell quantification by analyzing biological data in 3D rather than as individual 2D layered images.
Position Type	Intern
Training Plan	The undergraduate students at Kean University applying for this position are involved in a course series at Kean entitled Research First Initiative, which includes a research methods class, as well as support for incorporating a mentored research project for course credit. Additionally, “STEM: 1500 Introduction to Scientific Computing,” is a required course for students both in our computational science and biotechnology program, and our calculus sequence includes additional training in numpy/python. All students in the program will have daily stand-up meetings through the duration of the project, as well as weekly individual and group meetings to cover training needs in a more in-depth level when needed. Zoom and Slack will be used for virtual communication as needed to maintain social distancing while Covid-19 based restrictions on meetings are still in effect. Our mentoring plan also includes helping students prepare posters and talks for research conferences. Students that conduct research at Kean are supported with $1,000 to support conference attendance. One conference where our VR analysis tool will be presented is the virtual “Annual Drosophila Conference,” organized by the Genetics Society of America. Our training plan also includes responsible conduct of research training. Research students must pass a course which cover different aspects of ethics in research, “STEM: 1903 – Research Methods” prior to joining a research project. The format of this course is discussion based and covers topics such as plagiarism, responsible authorship, research misconduct, and proper data recording. Bianca Ortega, one of the undergraduates recruited for this position, has already made exceptional progress on creating a “proof of concept” tool through her current Apprentice XSEDE position. Continuing this project as an internship will give her the opportunity to learn how to automate and improve the run time efficiency of her current tool.
Student Prerequisites/Conditions/Qualifications	Students applying for this position are enrolled in the STEM honors program at Kean University, where the curriculum is designed to merge the teachings from various fields including biology and computational science. The undergraduate students that have been recruited to apply for this position major in Computational Science. Combined, the students have strong backgrounds in biology by completing General Biology I &II, statistical programming background by completing an R-based computer programming course, and basic computer programing skills from computer science courses. The students also completed two semesters of hands-on research experience in either Computational Biology in the Niepielko Lab or developing Unity based visualizations tools in the Joiner Lab. Together, our interdisciplinary team will have the Biology and Computational qualifications and experience necessary for the successful completion of this project that requires collaboration.
Duration	Summer
Start Date	06/01/2021
End Date	08/16/2021