Actuarial scientist publishes in Nature

It is a breakthrough in understanding the mammalian brain. Using a new ground-breaking method, researchers from the Department of Chemistry at the University of Copenhagen have closely studied an enzyme (V-ATPase) and discovered that it switches its activity on and off at completely random intervals, which contradicts previous knowledge about a vital process in the brain. That is why the research appeared on the cover of Nature.

Two mathematicians are co-authors of the article: Associate Professor Jesper Lund Pedersen from the Department of Mathematics, and Peter Johnson, a former postdoc at the department, currently employed at the University of Manchester. They both work with applied probability theory, mainly within financial mathematics and insurance mathematics.

How did an insurance mathematician end up in a biochemical research project?

- It was thanks to a Villum Experiment grant I got a few years ago. Villum called the recipients to some network meetings, and there I met Professor Dimitrios Stamou from the Department of Chemistry. We talked about our current research and Dimitrios could use some mathematicians for his project.

Dimitrios Stamou was also happy about the chance meeting:

- Jesper and Peter's efforts for the project were extremely important. They invented a new static method to analyze the project's data, he says. 

The algorithm finds changepoints in the noise

The first author of the article in Nature, Eleftherios Kosmidis from the Department of Chemistry, explains that the major breakthrough in their study was the finding that single mammalian-brain V-ATPase enzymes do not work continuously in time but rather stochastically switch between three different stages.

- To be able to reach this conclusion, we used fluorescence microscopy to monitor the signal produced by fluorescent molecules that report the activity of the enzyme by changing their total fluorescence output based on the mode that the enzyme is currently at. This was done in parallel, for thousands of enzymes at the same time.

- However, the nature of light introduces noise into our system. Jesper and Peter developed a Bayesian technique that was able to systematically and reliably “read” the intrinsic noise of the light collected and thus was able to deconvolve which fluctuations in the signal were due to random noise and which were due to the enzyme changing modes.

- This event detection technique assisted in extracting valuable physical properties of these newly uncovered modes of function such as switching frequency and mode probabilities, explains Kosmidis.

Jesper Lund Pedersen calls it change-points:

- When the enzyme or the cell starts to open, the data really comes. However, a movement can be random noise. We have programmed an algorithm – or estimation method – that can identify change points hidden inside the noise.

- The fact that the enzyme opens and closes, and then opens again... Few people can create such an algorithm with several change points. Most people have developed an algorithm with one or two points, but we can do it with many, says Jesper.

Will optimize the code

Jesper Lund Pedersen and Peter James are still involved in the project. After the initial experiment, which led to the article in Nature, the chemists will go up on a larger scale. They must measure several things in longer time series.

- The amount of data will grow a lot, which is why the algorithm's code must be "speeded up". The code must be optimized so that the computer can calculate faster, explains Jesper.

Eleftherios Kosmidis is looking forward to continuing the project:

- It was a fantastic and fruitful collaboration that can continue towards new and exciting directions, says Kosmidis.

---

Read more

Major discovery about mammalian brains surprises researchers

A proton-pumping enzyme in the brain switches between modes - article from Nature 23-11-2022