Monday 17 April 2023

ERC panel and grant links

Grant awards








Wednesday 23 February 2022

ARTICLE: Cellular energy budgets and antimicrobial resistance

Dynamic Boolean modelling reveals the influence of energy supply on bacterial efflux pump expression
Ryan Kerr, Sara Jabbari, Jessica MA Blair, Iain G Johnston
Journal of the Royal Society Interface 19 20210771 (2021)

Antimicrobial resistance or AMR is a major global health issue, with disease-causing organisms like bacteria acquiring resistance to the drugs we use to kill them. One way that bacteria acquire this resistance is through so-called efflux pumps -- cellular machinery that removes chemicals like drugs from inside the bacterium. Bacteria produce these pumps when faced with drug treatments, but not all cells produce the same amount or at the same time. Understanding this variability could help the theory behind future treatments.

After finding that the available "energy budget" influences the behaviour of cellular programs, we asked whether energy variability could be a cause of these differences. Using lots of diverse experimental observations, we built a theoretical model of the signals that tell a bacterium to produce efflux pumps in response to sensing a drug, with a new and simple way of accounting for how energy affects these signals. We then simulated this model in a computer to see how model cells with different amounts of available energy (as we see in real bacterial populations) behaved.

We found that differences in cellular energy budgets can have a profound effect on when, and how much, efflux machinery is produced. This variability builds on the natural randomness of the system, leading to several interesting results: energy changes the dynamics of how signalling programs work in the cell, alters timescales, and affects the "priming" of a population of cells to anticipate future stress. The approach we developed is quite general and can be used to explore energy influence on any other cell programs and signals too.

Including ATP, an important cellular energy currency, in models of how bacteria express efflux machinery helps us understand how cell-to-cell differences in energy budget may influence AMR.

ARTICLE: Removing mutant mtDNA from cells

2-Deoxy-D-glucose couples mitochondrial DNA replication with mitochondrial fitness and promotes the selection of wild-type over mutant mitochondrial DNA
Boris Pantic, Daniel Ives, Mara Mennuni, Diego Perez-Rodriguez, Uxoa Fernandez-Pelayo, Amaia Lopez de Arbina, Mikel Muñoz-Oreja, Marina Villar-Fernandez, Thanh-mai Julie Dang, Lodovica Vergani, Iain G Johnston, Robert DS Pitceathly, Robert McFarland, Michael G Hanna, Robert W Taylor, Ian J Holt, Antonella Spinazzola
Nature Communications 12 1 (2021)

This is an exciting one! As we've discussed before, mitochondrial DNA (mtDNA) molecules exist in large populations in our cells, encoding vital machinery. Devastating diseases can result when a high proportion of a cell's mtDNA molecules are mutated, but cells can deal with a low proportion of mutant mtDNA. So, it'd be great if we had a way to decrease the proportion of mutant mtDNA in cells -- below the threshold for disease.

Perhaps we do! We recently played a supporting role in a project with Antonella Spinazzola and Ian Holt, looking at what happens when cells containing a mixture of mutant and normal mtDNA are treated with chemicals. They found that a molecule called 2DG (for short) slows down the replication of mutant mtDNA in cells. As mtDNA is constantly replicating, this preferential inhibition of mutant means that normal mtDNA comes to dominate the cellular population over time. We showed this population shifting over time in a variety of human cell lines and growth media, including several chosen to model in vivo behaviour.

The project showed that 2DG compromises mitochondrial respiration much more in mutant than in wildtype mitochondria. This is likely why mutant replication was so challenged -- poorly functioning mitochondria are less likely to replicate. Restricting glutamine and glucose together had the same effect (though is perhaps harder to achieve in a therapeutic context). 2DG is in trials for epilepsy treatment, so may represent a path to new therapies addressing mtDNA disease. There's a press release here with some more commentary.

ARTICLE: Social networks of plant mitochondria

Network analysis of Arabidopsis mitochondrial dynamics reveals a resolved tradeoff between physical distribution and social connectivity
Joanna M Chustecki, Daniel J Gibbs, George W Bassel, Iain G Johnston
Cell Systems 12 419 (2021)

We recently spent some time looking at a long-standing question in plant cell biology -- why do mitochondria move the way they do? Plant mitos look for all the world like cars in a city, moving along highways and speedily getting from place to place. We combined laser microscopy, video analysis, physical modelling, and network science to explore what benefits this motion might bring to the cell. It turns out, it allows mitochondria to have social lives! Through the "social network" of encounters between moving mitochondria, beneficial exchange of contents can occur, while their motion allows the cell to keep its population well spread. Here's a blog article from Jo explaining things more!


... and also check out Jo's beautiful site!

Mitochondria are in yellow in the microscopy image; their "social network", describing their encounters, is overlaid in white. Cover of the month's Cell Systems issue.

ARTICLE: Stochastic fantasy combat!

Optimal strategies in the Fighting Fantasy gaming system: influencing stochastic dynamics by gambling with limited resource
Iain G Johnston
European Journal of Operational Research DOI 10.1016/j.ejor.2022.01.039 (2022)

Here's a more unusual one. Fighting Fantasy gamebooks, hugely popular in the 80s and resurgent now, are "games in a book". The reader/player chooses their path through the book's many sections, overcoming challenges, fighting monsters, and hopefully succeeding in their quest.

The combat system is quite interesting. The player and their opponent both have "stamina" -- like the health bar in a video game. The player rolls dice to determine the strength of one of their attacks, and rolls again for the opponent. Whoever has the higher strength inflicts some stamina loss on the other. The combat proceeds through rounds like this until someone's stamina reaches zero.

Phrased like this, the player has no agency and the system is quite easy to solve (ie, work out the probability of winning a given fight). But there's another factor. The player (not the opponent) also has some "luck", describing how lucky they are. Testing luck involves rolling two dice: if the sum is less than or equal to the player's luck score, they are luck, otherwise they're unlucky. If they win an attack round, they can choose to test their luck, and if luck they deal more damage to their opponent. If they lose an attack round, they can also test their luck, and will take less damage if lucky. If they're unlucky, the outcome is negative: they do less, and take more, damage than if they'd not tested.

A bit more complicated, but still possible to solve. But here's the rub. Every time you test your luck, your luck score decreases -- whether you're lucky or unlucky. So as you test your luck more and more, you are less and less likely to get a positive outcome. The question is -- when is it a good idea to use a luck test in combat?

To answer this we used an approach called dynamic programming. We first considered all the ways combat can end -- with someone having no stamina left. We next considered every state of the combat that could lead to one of these end states, and calculated the probability of each possible outcome in the case where the player chose to test their luck and the case where they didn't. We then considered all states of combat that led to these states, and so on, multiplying probabilities as we went to calculate the overall probability of victory from any state given any luck strategy.

We found that judicious use of luck can dramatically increase victory probability in some cases, particularly when player and opponent statistics are unbalanced. There are some general rules -- for example, no matter how low your luck, you should always test if you are otherwise about to die. We used some tools from statistics to distil the complex set of detailed optimal strategies into more general principles to follow. We also connect to more real-world questions, like cheating and espionage, where a "lucky" outcome can be beneficial -- but an unlucky one can be disastrous, and the more you test your luck the more likely you are to be unlucky!

ARTICLE: Corals to crops -- how life protects the plans for its cellular power stations

Avoiding organelle mutational meltdown across eukaryotes with or without a germline bottleneck
David M Edwards, Ellen C Røyrvik, Joanna M Chustecki, Konstantinos Giannakis, Robert C Glastad, Arunas L Radzvilavicius, Iain G Johnston
PLoS Biology 19 e3001153 (2021)

(this text is from a press release about the article)

An international team of researchers led by the University of Bergen has uncovered how organisms from crops to corals may avoid deadly DNA damage during evolution.

Our cells, and those of animals, plants and fungi, contain compartments that produce chemical fuel. These compartments contain their own DNA, which stores instructions for important cellular machinery. But this so-called oDNA (organelle DNA) can become mutated, corrupting the instructions and preventing cells making enough energy.

In humans and some other animals, a process called the “bottleneck” allows some offspring to inherit less mutated oDNA. This process needs mothers’ egg cells to develop early, like in humans, where a human girl is born with all her egg cells already formed. But other organisms, from plants to fungi, don’t develop these cells early – their flexible body plans mean that eggs are not “set aside” early in development.<\p>

“We wanted to know how these organisms might avoid inheriting mutations without a human-like bottleneck,” said Ellen Røyrvik, a geneticist on the research team, based at UiB.

The scientists used mathematical modelling to show that a process called gene conversion – the controlled overwriting of DNA – could in theory allow some offspring to inherit less mutant oDNA without requiring a bottleneck. Using genome data, they found machinery controlling this process in plants and fungi, but also in soft corals, sponges, and algae – all organisms without fixed body plans. They also found that this machinery was most active in the parts of plants that will end up producing the seeds of the next generation, suggesting that it is indeed used to allow some offspring to inherit fewer mutations.

Organisms without fixed body plans (including octocorals, sea pens, sponges, plants, and fungi) and with fixed body plans (including humans and many animals) may use different strategies to avoid the buildup of damage in their cellular "power stations." CREDIT: Gemma Lofthouse

“Taken together, it looks like organisms without a fixed body plan – plants, fungi, corals, sponges, algae – may have adopted gene conversion to deal with oDNA mutations,” said Iain Johnston, an associate professor in the Mathematics Institute at UiB, who led the research. “Humans and other animals can develop egg cells early and use a bottleneck; other organisms can use gene conversion instead.”

Going forward, the team plans to explore how this overwriting of oDNA causes other issues in the organisms that use it – including crop plants, where it can cause sterility. They are also exploring the broader question of why these compartments contain oDNA at all, given the risk of mutational damage.

Friday 23 April 2021

ARTICLE: How does tool use evolve in animals?

Data-Driven Inference Reveals Distinct and Conserved Dynamic Pathways of Tool Use Emergence across Animal Taxa, iScience 23 101245 (2020)

There are some wonderful examples of animals using tools. Octopuses block oyster shells open with coral; boxer crabs wave captive anemones for defense and food capture; captive dolphins use feather to wipe clean their aquarium windows. A while ago, we saw this excellent infographic in National Geographic, and got interested in these data. Do different families of animals learn to use tools in completely different ways? Or are there some general ("universal") principles behind how animals learn to use tools?

There is, of course, a big and fascinating literature on tool use, but we found rather few studies attempting a quantitative comparative analysis across bilaterian animals. To address other evolutionary questions, we've developed HyperTraPS (hypercubic transition path sampling), a statistical approach for learning the "pathways" of evolutionary processes. That is, which events occur before and after which other events in an evolving system? Does feature A always evolve before feature B? We used HyperTraPS to ask about the orderings with which different types of tool use appeared in animals. For example, do animals always learn to "poke" before they learn to "dig"? Do all animals learn tool use in the same order, first A then B then C..., or does it vary across species?

We found some answers that we think are quite interesting. There seem to be some quite deep similarities across animal species in how tool use evolves. Types of tool use like "affixing" and "throwing" are almost universally acquired early; types like "cutting" and "symbolising" are acquired late and rarely, only by primates. The environment and animal family influences the structure of these pathways: aquatic organisms seem to discover "waving" tool use relatively early, for example, and primates discover tools that "block" relatively late.

(A) The inferred pathways of tool use emergence across animals. The size of a blob gives the probability that that mode of tool use (on the horizontal axis) is acquired at that stage (on the vertical axis) of a species' discovery of tool use types. (B) Sample evolutionary pathways of tool use, with individual animal lineages illustrated at the positions corresponding to the modes of tool use they have discovered.

Of course, there's a lot of uncertainty about any analysis like this. Are we talking about wild or captured animals? What if we just haven't observed some types of tool use? We attempted to address several such questions with our analysis and showed that our overall results were quite robust with respect to these uncertainties. HyperTraPS fully describes the uncertainty in its outcomes, helping interpretability. We hope that our results help at least to suggest some possible principles and points for further investigation in this fascinating topic. You can read more in iScience here.