Hi! I am a postdoctoral researcher in the research group of Michael Stumpf at the University of Melbourne. My work focusses on mathematical and statistical methods in biology, particularly applying Bayesian methods and Machine Learning to model cells. I recently completed my PhD at the University of Edinburgh under the supervision of Guido Sanguinetti and Ramon Grima. Prior to that I was a research assistant in Jakob Macke’s research group, working on likelihood-free inference for biophysical models of neurons.

Research

Randomness in biology: My research focuses on stochastic models in biology. It is nowadays well-established that random events occur at all levels in biology, from individual genetics (what eye colour would my children have? could they inherit genetic diseases? ) to populations and ecosystems (how do species evolve or become extinct? ). Some of these can have significant impact on our lives, such as what are my chances of getting cancer? and when is the next pandemic going to happen?. Many of these processes start at the cellular and molecular level, and one of my research topics is understanding the role and importance of randomness at this stage.

Stochastic modelling: My current work centers on the Chemical Master Equation, which provides a rigorous way of describing many biological systems subject to randomness. The Chemical Master Equation is notoriously tricky to handle, and I am interested in finding ways to approximate/simplify it - this will help researchers make better predictions with less effort. To this end I focus on automated computational tools based on statistical learning and inference.

Inference: Since observing biological organisms at cellular level is difficult, my work involves a lot of mathematics and statistics to extract information from what we can measure. I am interested in developing and adapting inference methods that are scalable and easy to use, so that researchers without a degree in Machine Learning (or a spare grad student) can feel confident about applying them. I am particularly interested in Likelihood-Free Inference (e.g. Approximate Bayesian Computation) as many real-life problems in biology are too difficult to solve using classical likelihood-based approaches.

Further Reading: For an accessible introduction to the Chemical Master Equation, its approximations and some relevant inference methods I highly recommend the following resources:

• D. Schnoerr, G. Sanguinetti, R. Grima: Approximation and Inference for Stochastic Biochemical Kinetics - a Tutorial Review
• Darren Wilkinson: Stochastic Modelling for Systems Biology

If you want to know more about my research, please check out my contact details!

Publications

2024

• K. Öcal, M.P.H. Stumpf. A Universal Formula Explains Cell Size Distributions in Lineages, preprint
  [paper]
• L. Ham, M.A. Coomer, K. Öcal, R. Grima, M.P.H. Stumpf. A Stochastic vs Deterministic Perspective on the Timing of Cellular Events, Nat. Commun. 15(1)
  [paper] [code]

2023

• K. Öcal. Incorporating Extrinsic Noise into Mechanistic Modelling of Single-Cell Transcriptomics, preprint
  [paper] [code]
• K. Öcal, G. Sanguinetti, R. Grima. Model Reduction for the Chemical Master Equation: An Information-Theoretic Approach, J. Chem. Phys. 158(11)
  [paper] [code]

2022

• A. Sukys, K. Öcal, R. Grima. Approximating Solutions of the Chemical Master Equation Using Neural Networks, iScience 25(9)
  [paper] [code]
• K. Öcal, M.U. Gutmann, G. Sanguinetti, R. Grima. Inference and Uncertainty Quantification of Stochastic Gene Expression via Synthetic Models, J. R. Soc. Interface 19(192)
  [paper] [code]

2020

• P.J. Gonçalves, J.-M. Lueckmann, M. Deistler, M. Nonnenmacher, K. Öcal, G. Bassetto, C. Chintaluri, W.F. Podlaski, S.A. Haddad, T.P. Vogels, D.S. Greenberg, J.H. Macke. Training Deep Neural Density Estimators to Identify Mechanistic Models of Neural Dynamics, eLife 9
  [paper] [code]

2019

• K. Öcal, R. Grima, G. Sanguinetti. Parameter Estimation for Biochemical Reaction Networks Using Wasserstein Distances, J. Phys. A 53(3)
  [paper] [code]

2017

• J.-M. Lueckmann, P.J. Goncalves, G. Bassetto, K. Öcal, M. Nonnenmacher, J.H. Macke. Flexible Statistical Inference for Mechanistic Models of Neural Dynamics, Adv. Neural Inf. Process. Syst. 30 30
  [paper] [code]

Software

A lot of my research involves scientific programming, for which I use Python, Stan, Julia and C/C++. While the documentation is sparse in some places, I always welcome issues and pull requests on GitHub and can be contacted for any questions. All my code is intended to be freely used by others and is currently maintained. Apart from the code linked in the above list of publications, I have worked on the following standalone project:

FiniteStateProjection.jl

This Julia package implements Finite State Projection algorithms to solve the Chemical Master Equation numerically. Compatible with the broader Julia SciML ecosystem. Don’t forget to check out its sister package MomentClosure.jl, the original inspiration for this work!

Contact

I am active on Github.

You can contact me at firstname.lastname@unimelb.edu.au (lastname starts with an ‘o’). Feel free to get in touch if you have any questions concerning my research or the software I’ve been working on, or if you wish to contact me for outreach activities.