Tracking Single Cells Using Deep Learning

Ulicna, Kristina; Communications, Corporate

November 11, 2021

Tracking Single Cells Using Deep Learning

A video presentation from Kristina Ulicna at the AI microscopy symposium

Separation of cells based on their tracking status: A colourised binary mask of a time-lapse microscopy field of view of medium confluency with individual cells highlighted as survivors if they can be tracked since the initial movie frame (cyan), incomers if they migrated into the field of view throughout the movie (yellow) or mistracks if an error occurred in the automated trajectory reconstruction (red). Tracking_single_cells_using_deep_learning_teaser.jpg

AI-based solutions continue to gain ground in the field of microscopy. From automated object classification to virtual staining, machine and deep learning technologies are powering scientific breakthroughs while helping microscopists streamline analysis.

Image: Separation of cells based on their tracking status: A colourised binary mask of a time-lapse microscopy field of view of medium confluency with individual cells highlighted as survivors if they can be tracked since the initial movie frame (cyan), incomers if they migrated into the field of view throughout the movie (yellow) or mistracks if an error occurred in the automated trajectory reconstruction (red).

Speaker: Kristina Ulicna

Kristina holds a BSc in Biomedical Sciences from King’s College London and is currently working on her PhD at University College London with Alan Lowe and Guillaume Charras. Her PhD research utilizes deep learning to track single cell heterogeneity within non-cancer and cancer cell lines. Kristina’s work in this area was recently accepted to Frontiers in Computer Science: www.frontiersin.org/articles/10.3389/fcomp.2021.734559/abstract

She also recently completed an internship at Microsoft Research Cambridge as an AI research scientist and Forbes Slovensko shortlisted her as a 2021’s under-30 laureate.

In this presentation, Kristina covers the tools she has developed to automatically detect cell nuclei using deep learning, track each nucleus over time, and visualize the complete cell lineage tree for individual cells. These tools allow her capture the cellular state over multiple generations and to examine whether cell cycle durations are inheritable.

More about her work: https://github.com/KristinaUlicna

Key Learnings

Using multiple deep learning models and bTrack algorithms cell segmentation, cell state classification and single cell tracking can be fully automated
Full single cell lineage tracing is possible and allows questions regarding heritable traits can be examined
Wild-type MDCK cells have a large range of cell cycle durations and that this range may be maintained across lineages

Register to view the video presentation

First Name*

Last Name*

Email*

Position*

Zip*

Country/Region*

Questions or Comments

I am interested in receiving information about products and services, virtual and in-person events such as workshops and webinars and educational content from Leica Microsystems GmbH, their representatives and other Danaher companies.
I understand that I may be contacted by phone or email and consent to such communication. *

Yes

No

By clicking on the SUBMIT button, I confirm that I have reviewed and agree with Leica Microsystems GmbH Terms of Use and their Privacy Policy. I also understand my privacy choices as they pertain to my personal data, as detailed in the aforementioned Privacy Policy under ‘’Your Privacy Choices’’.

mtCaptcha