CYTO Virtual Interactive 2021 Oral Presentation - TrackSOM: Mapping Immune Response Dynamics Through Temporal Clustering and Tracking of Single-Cell Cytometry Data

  • Register
    • Visitor - Free!
    • Bronze - Free!
    • Silver - Free!
    • Gold - Free!
    • Platinum - Free!
    • Community Administrator - Free!
    • ISAC Staff - Free!


Temporal high-dimensional cytometry data, which captures the characteristics of cells in biological samples collected over a period of time, are routinely collected to study the complex and dynamic immune response. Mapping out the immune response and relating the differences therein with clinical outcomes are crucial to understand how it operates and identify intervention opportunities. In cytometry, computational algorithms which assist in the mapping of the dynamics of immune responses by objectively delineating cellular populations and tracking their changes over time are lacking. We present TrackSOM, a temporal clustering and tracking algorithm which automatically identifies cellular populations in temporal high-dimensional cytometry data and tracks their development over time. We applied TrackSOM to elucidate the evolving immune response in both the bone marrow and brain of West Nile Virus-infected mice. In the bone marrow, TrackSOM tracked the progressive up-regulation of SCA-1 expression by several clusters and demonstrated its capacity to uncover the heterogeneity within the response of different cell types to infection. In the brain, TrackSOM uncovered a gradual increase and decrease in the proportion of infiltrating macrophages and microglia respectively, and showcased its ability to elucidate antiviral response to the WNV encephalitis. Additionally, TrackSOM was used to assess immune activation over time in the blood of COVID-19 patients, and revealed the activation of innate, adaptive, and myeloid cells during the acute and convalescent phases. TrackSOM provides insight into the dynamic development of immune response and paves the way for the mapping and extensive investigation of immune responses in wide range of diseases.



Givanna Putri
PhD Candidate
University of Sydney

Givanna Putri is a PhD scholar from The School of Computer Science at The University of Sydney, supervised by Dr. Mark Read and Prof. Irena Koprinska, and a member of The Charles Perkins Centre. She received her Bachelor degree (with first class Honours) in Information Technology from The University of Sydney in 2015, and worked professionally as a software developer between 2015-2017. Her current PhD research was funded by The Australian Government Research Training Program Scholarship and The University of Sydney Alumni Scholarship, investigating the use of temporal clustering and tracking algorithm to reveal the development of immune response over time.
During her PhD, she developed machine learning techniques which explicitly cluster and track cellular populations development in time-course high dimensional cytometry data. This included the TrackSOM algorithm which was recently used to profile SARS-CoV-2/COVID-19 patients, and the ChronoClust algorithm which publication on the Knowledge Based System journal won the Dolby scientific paper competition in 2019. Additionally, she also developed a novel benchmarking methodology ParetoBench which uses the Pareto Front framework to fairly compare the performances of clustering algorithms using several performance metrics.
She was awarded The Student Travel Award from the International Society for Advancement of Cytometry (ISAC) and The Postgraduate Research Support Scheme from The University of Sydney to present ChronoClust at the 34th Congress of the International Society for Advancement of Cytometry (CYTO) conference. She is extremely interested in inter-disciplinary research, specifically in using Machine Learning and Artifical Intelligence to advance life sciences.

CMLE Credit: 1.0


Oral Presentation
Recorded 06/07/2021
Recorded 06/07/2021
CMLE Evaluation Form
11 Questions
11 Questions CMLE Evaluation Form
Completion Credit
1.00 CMLE credit  |  Certificate available
1.00 CMLE credit  |  Certificate available