Skip to content
View KristinaUlicna's full-sized avatar

Highlights

  • Pro

Block or report KristinaUlicna

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
KristinaUlicna/README.md

Welcome to Kristina's GitHub Page! πŸ‘‹

πŸ’‘ Let's get straight to the point: Check out my CV πŸ“„

My name is Kristina Ulicna, PhD & I'm a research associate at The Alan Turing Institute πŸ‘©β€πŸ’». I obtained my PhD degree from University College London (UCL) πŸŽ“ as part of the London Interdisciplinary Doctoral (LIDo) Bioscience Programme, the largest BBSRC-funded Doctoral Training Partnership in the UK πŸ‡¬πŸ‡§. I'm a computational biologist focusing on the interplay between single-cell trajectories reconstruction πŸ‘£, representation πŸ”’ & interpretation πŸ“Š via AI-driven bioimage analysis tool development πŸ’».

Github Linkedin Twitter Gmail Outlook ResearchGate GoogleScholar

Me, a Computational Biologist

LIDo_Int

I'm a trained biomedical scientist turned computational biologist. After obtaining my BSc degree in Biomedical Science @ King's College London, I started my PhD training with the London Interdisciplinary Doctoral (LIDo) Programme @ University College London (UCL). I completed my PhD under the guidance of Drs Alan Lowe (The Alan Turing Institute 'AI for Science' Fellow, UCL) & Guillaume Charras (London Centre for Nanotechnology, UCL).

Note: If you are a non-scientist interested in what I do, before you get overwhelmed by the scientific jargon & terminology, check out my #DeepTree🌳 Tweetorial Thread 🧡 for an easy-to-digest breakdown of my research for lay audience. In case you're not scared of the technical terms, don't hesitate to check out my papers in the Citation section find out how to properly cite our work, or read the brief description below:

The overall theme of my PhD thesis was quantitative labelling of single-cell trajectories in time-lapse microscopy. That said, the work could be easily split into two parts:

  1. Trajectory reconstruction & lineaging
    • Co-developed a robust, supervision-free, deep learning-based cell tracking pipeline for deep lineage analysis of live-cell microscopy 2D cell lines data
    • Analysed multi-generational lineage trees of >20k single-cell trajectories to interpret proliferation characteristics predisposing cells to fast divider rates
  2. Track representation & interpretation
    • Generated an explainable AI model to learn dynamic image representations & interpretable latent space features to map similarities of cell cycle continuity
    • Transformed a sequence of image repre-sentations into an unsupervised trajectory annotation, classifying cell cycle phases & quantitative confidence scoring over time

Me, a Research Scientist

Past Achievements

I'm a trained cancer biologist with a 1st class Biomedical Science degree from King's College London. I'm a practical & detail-oriented researcher with a unique combination of wet-lab & dry-lab skills which I gained through various internship experience in both academia & industry. To point out a few, I interned in the famous Robert Weinberg's Lab at Massachusetts Institute of Technology (MIT) in Cambridge, MA, USA where I worked on the identification of a novel tumour suppressor gene. Most recently, Microsoft Research Cambridge Inner Eye group welcomed me to their team for 6-month PhD internship, which I spent contributing to the SOTA solutions of a (bio)medical image analysis-oriented Kaggle challenge in collaboration with the Human Protein Atlas database founders.

Lab_pic

Present Interests

Now as a PhD graduate, I'm interested in deep learning-based representation of single-cell trajectories in cell populations. I use a combination of time-lapse microscopy movies to enhance my passion for big data science, deep learning & computational cell tracking to understand morphological features influencing single-cell cycling heterogeneity. I'm actively developing AI-driven tools to answer my PhD thesis project goal - quantitatively follow and temporally label cell cycle trajectories from cell's early life to cell division. To do so, I combine feature handcrafting tools with variational autoencoders (VAE) approaches to encode the single-cell sequences throughout their lifetime for cell trajectory reconstruction. I'm clustering these data to find (dis-)similarities between the individual cells. Using time-sequence analysis methods, I shortlist the regions of the sequence data to identify patterns leading to pre-determination of cell cycle lifetimes.

Future Outlooks

I'm always in search for interesting full-time Research Scientist / Computational Biologist positions at the interface of Machine Learning & Biological / Biomedical Research. If my work sparked some interest in you, do not hesitate to get in touch! Don't forget to check my up-to-date CV for a better overview of my skillset, work / teaching experience & professional interests, or write me an email / message me directly using the social media links in the panel above. Always happy to hear from you!


Me, a Science Communicator

I love talking about my research to a variety of audiences! Whether you're an expert with 20 years worth of research experience or a lay listener fascinated by biomedical science, you're at the right place... Let me list a few resources you may be interested in having a look at to learn more about my research interests: use this traffic light system to navigate the links based on your level of expertise:

LIDo_Int

  • 🟩 : lay audience with interest to learn more
  • 🟧 : scientist / technologist from outside of this field
  • πŸŸ₯ : expert 'singlecellologist' who enjoys the niche jargon πŸ€“)
  1. 🟒🟩🟒 In this Twitter thread 🧡, I introduced the objective of my PhD research project & how we tackled the single-cell heterogeneity in cell populations by developing super-cool live-cell microscopy & computational tools... πŸ§ͺ

  2. 🟒🟩🟒 I spoke about my research at the PyLadies Meetup πŸ‘©β€πŸ’» where I delivered a talk about how Python (the programming language, not the snake 🐍!) can serve as an incredibly useful tools to (not only computational) biologists... πŸ’»

  3. 🟠🟧🟠 This short 12-second LinkedIn movie 🎬 illustrates how much information πŸ“Š regarding cell cycle control one may squeeze out from a sequence of microscopy images using machine (deep) learning... πŸ”¬

  4. 🟠🟧🟠 This visually-appealing Twitter thread 🧡 summarises the key outcomes πŸ”‘ of the "Introduction to Deep Learning" MasterClass (with Youtube link to the talk) which I recently delivered at the UCL Cancer Domain Early Career Network... πŸ‘©β€πŸ«

  5. πŸ”΄πŸŸ₯πŸ”΄ I describe the scientific details of my PhD thesis research project in my CellComp repository where I explain the background, objectives, methodology & key results of my doctoral thesis project... πŸ‘©β€πŸ”¬

  6. πŸ”΄πŸŸ₯πŸ”΄ And last but not least, our recent publications to be found at Frontiers journal website or at the bioRxiv preprint repository should give you all the details you need to know about what we've done in the lab & beyond. ➑️ We also encourage you to jump to my DeepTree repository if you'd like to learn even more... 🌳


What keeps me busy? πŸ‘©β€πŸ’» Live projects 🚧

  • πŸ‘©β€πŸ’» Co-developing graph representation analysis for connected embeddings (GRACE) for automated object identification of structural patterns in (bio-)imaging datasets
  • πŸ‘― I’m looking to collaborate on bioscience research ideas & projects involving AI;
  • πŸ’Ό I'm in search of interesting technical roles & job openings to enhance my skills;
  • πŸ’¬ Ask me about anything -> I am happy to answer your questions & help you out;
  • πŸ“« How to reach me: check the banners on top / bottom of this page!

Open-Source Projects 🚧

I've been contributing to a handful of open-source projects lately, mastering my software engineering skills & how to improve my programming practices. Check out the projects below or have a direct look at my contributions at some of the repositories I helped to build by visiting my GitHub repositories link πŸ”— for more details (links coming soon!).

Picture 1 Picture 2 Picture 2

Languages & Tools πŸ”§

Kristina's GitHub Stats

Getting in Touch πŸ’¬

If you're interested in more details about:

  • My professional background & research experience, have a look at my CV πŸ“„ or choose to click on the links below πŸ‘‡

Github LinkedIn Twitter Gmail Outlook ResearchGate GoogleScholar


Citations

LIDo_Int

More details about my recent publications focussed on single-cell tracking approach from time-lapse microscopy data can be found in my DeepTree repository or in the following publications:

Learning dynamic image representations for self-supervised cell cycle annotation
Ulicna K, Kelkar M, Soelistyo CJ, Charras GT & Lowe AR
ICML Computational Biology Workshop (2023) Links: Workshop | bioRxiv

Perspective: Machine learning enhanced cell tracking Soelistyo, C.J., Ulicna, K. & Lowe, A.R. Frontiers in Bioinformatics, Expert Opinions in Computational Bioimaging (2023) Links: Frontiers

Convolutional neural networks for classifying chromatin morphology in live cell imaging
Ulicna K, Ho LTL, Soelistyo CJ, Day NJ & Lowe AR
Methods in Molecular Biology, Springer Nature Protocols (2022) Links: Springer

Automated deep lineage tree analysis using a Bayesian single cell tracking approach
Ulicna K, Vallardi G, Charras G & Lowe AR
Frontiers in Computer Science, Computer Vision: Methods & Tools for Bioimage Analysis (2021)
Links: Frontiers | bioRxiv

Pinned Loading

  1. DeepTree DeepTree Public

    Jupyter notebooks for deep lineage analysis of single-cell heterogeneity and cell cycling duration heritability.

    Jupyter Notebook 30 7

  2. CellComp CellComp Public

    Python Scripts to my Cell Competition project contribution. Looking at pure WT cell populations to identify patterns of CSC.

    Jupyter Notebook 6 3

  3. BayesianTracker BayesianTracker Public

    Forked from quantumjot/btrack

    Bayesian multi-object tracking

    C++ 4

  4. CellTracking CellTracking Public

    Forked from quantumjot/CellTracking

    Cell tracking using TensorFlow and BayesianTracker

    3 1

  5. sequitr sequitr Public

    Forked from quantumjot/sequitr

    Microscopy image processing with TensorFlow

    Python 2

  6. unet_segmentation_metrics unet_segmentation_metrics Public

    Forked from lowe-lab-ucl/unet_segmentation_metrics

    Metrics for determining segmentation accuracy

    Jupyter Notebook 3