Icon Legend

This session is not in your schedule.

This session is in your schedule. Click again to remove it.

Biomedical Engineering Education

Biomedical Engineering Education - Poster Session B

Poster E9 - Decentralized crowd-sourced image-based biokits for distributed spatial biology and discovery

Thursday, October 24, 2024

2:45 PM – 3:45 PM EST

Location: Exhibit Hall E, F & G

Presenting Author(s)

Samyukta Iyer (she/her/hers)

Biomedical Engineering Student Researcher
Georgia Institute of Technology
Marietta, Georgia, United States

Co-Author(s)

Nicholas Zhang

Graduate Researcher
Georgia Institute of Technology, United States

Primary Investigator(s)

Ahmet F. Coskun

Bernie Marcus Early-Career Professor
Georgia Institute of Technology, United States

Introduction: The current conventional academic research model involves a small group of collegiate researchers spending significant effort on designing and performing the experiments in an expensive, central laboratory, and subsequently conducting data analysis in a distributed fashion, often resulting in time-consuming bottlenecks due to resource availability and technical skill limitations.
Thus, crowdsourcing is an emerging approach to rapidly and simultaneously engage individuals in distributed data collection from diverse and otherwise excluded potential researchers, creating more opportunities for innovative research outcomes and critically, reducing the logistical and financial barriers to biomedical engineering research for younger students.
This project tested how academic research experiences can become more accessible to inexperienced students from traditionally marginalized/minority backgrounds, using novel, low-cost tools, such as 3D-printed smartphone-compatible microscopes and a crowdsourced distributed research model. Here, the biokit, a low-cost imaging system, and computational analysis pipeline, was provided to students. To test the quality of research output from the crowdsourcing model, individual participants studied C. elegans locomotion and performed data collection and processing remotely using the biokit components. In addition, students used the biokit to conduct independent imaging studies to explore possible applications of this kit and crowdsourced research model, such as the morphology and size analysis of various objects. The success of this project shows that a crowdsourced research model is a viable means of conducting academic research as well as, equally importantly, expanding access to academic research opportunities, and can thus expand the future of research through expediting and diversifying inquiry and innovation.

Materials and

Methods: Comprehensive details, including CAD files, assembly instructions, and student-facing tutorial videos for usage of the biokit are found here: https://www.dropbox.com/scl/fo/ub3n2fsd6vfnalik8oyix/h?rlkey=24sxb75lrn0mf8sb1gg2tgvwc&dl=0.\

The biokit consisted of a 3D-printed, smartphone camera-compatible brightfield microscope, aspheric lens, LED light, and culture of N2 C. elegans. Initial plates included age synchronized L1 worms whereas the propagated plates were not age synchronized, resulting in a mixed age population. The 3D printed microscope was adapted from an open-sourced design in SOLIDWORKS and modified to be less resource-intensive to print and to include a lens adapter, a 3D printed Petri dish holder, and a customized light holder. High school age students were recruited through self-volunteering, with support from existing collegiate lab members.
Each student independently captured 1 to 5-minute videos of C. elegans using the biokit microscope and their individual smartphone on a daily basis, compiled in Google Drive over a 2 week period. Individual videos were then combined into a larger set in Dropbox. The set was divided out and computationally evaluated using Tierpsy Tracker for kinematics and locomotion features as a proof of concept for the biokit’s functionality and scientific effectiveness. Tierpsy Tracker was selected out of various other tracker options due to its batch processing and automatic analysis features, open source technology, results accuracy, and ease of use for novice student researchers.
Tierpsy Tracker results (approximately 4,500 features for 10 videos) were subsequently analyzed in Python using Min-Max normalization, dimensionality reduction using umap-learn, and Plotly.

Results, Conclusions, and Discussions: Over 100 individual videos were obtained, of which daily subsets were selected by students based on uniformity in imaging conditions and analysis parameters were determined. To effectively collaborate and proactively minimize human error in data collection and computational analysis, especially among beginner student researchers, frequent communication was also maintained, including regular checkups, small group meetings, and general team meetings.
Dimensionality reduction using Uniform Manifold Approximation and Projection (UMAP) and heatmap plots revealed that all given sets of videos analyzed by different researchers (R1-R4) were clustered closely and showed low variance, indicating results’ consistency despite crowdsourcing.
The study’s results show that robust and repeatable analytical results can be obtained asynchronously and independently, in a potentially faster manner than conventional research models. Critically, this is possible even when experimentation and analysis is completely planned and executed by beginner researchers with academics’ guidance, making this biokit model particularly valuable for expanding access to biomedical engineering education to underrepresented groups and empowering pursuit of post-secondary studies.
In conclusion, this study successfully leveraged the biokit model to train multiple student researchers with minimal to no background experience in either computational analysis or biological specimen research to independently collect and computationally analyze experimentally-relevant data as a team, gaining hands-on experience in every step of the scientific method and engineering design process, from developing experimental methods to writing reports. In the future, biokits could also be integrated with augmented reality (AR) and metaverse-based education, via creating immersive, individualized, and accessible lab simulations for students.

Acknowledgements (Optional): We thank Luigi Feriani for helping troubleshoot Tierpsy Tracker issues. AFC holds a Career Award at the
Scientific Interface from Burroughs Wellcome Fund and a Bernie-Marcus Early-Career Professorship. A. F. C.
was supported by start-up funds from the Georgia Institute of Technology and Emory University. Funding for
K.D. was provided by GT NSF-Simons Southeast Center for Mathematics and Biology and D.I.G. by the Dunn
Family Professorship. We thank Prof Hang Lu for providing the original worm cultures