2022 Winter Newsletter

Alys Clark (University of Auckland), Sara Loo (Johns Hopkins University), Fiona R. Macfarlane (University of St Andrews), and Thomas Woolley (Cardiff University).
  1. News – updates from: 
  2. People – Interview with Prof. Laura Kubatko, Chair of the upcoming Annual SMB meeting at Ohio State.
  3. Editorial – “Producing a good outreach talk”.
  4. Featured Figures – Highlighting the research by early career researcher, Alex Browning (University of Oxford) and highlighting the most downloaded paper from the Bulletin of Mathematical Biology in November 2022.

To see the articles in this issue, click the links at the above items.

Contributing content

Issues of the newsletter are released four times per year in Spring, Summer, Autumn, and Winter. The newsletter serves the SMB community with news and updates, so please share it with your colleagues and contribute content to future issues.

We welcome your submissions to expand the content of the newsletter.  The next issue will be released in May 2023, so if you would like to contribute, please send an email to the editors by the start of May 2023 to discuss how your content can be included. This could include summaries of relevant conferences that you have attended, suggestions for interviews, professional development opportunities etc. Please note that job advertisements should be sent to the SMB digest rather than to the newsletter.

If you have any suggestions on how to improve the newsletter and would like to become more involved and/or contribute, please contact us at any time. We appreciate and welcome feedback and ideas from the community. The editors can be reached at newsletter@smb.org.

We hope you enjoy this issue of the newsletter!

Alys, Sara, Fiona, and Thomas
Editors, SMB Newsletter



News Section

By Fiona MacfarlaneNews image

In this issue of the News section, we highlight the updates from the SMB Subgroups, Royal Society Publishing and the upcoming SMB Conference. Read on below.


SMB Subgroups Update

Immunobiology and Infection Subgroup

The Immunobiology and Infection subgroup, the Society for Mathematical Biology (SMB), and the National Institute of Allergy and Infectious Diseases (NIAID) are organizing a half-day workshop on Bridging multiscale modeling and practical clinical applications in infectious diseases.  The workshop will take place during the 2023 SMB meeting at Ohio State on July 19th from 1pm-5pm EST, with a networking opportunity to follow. Please save the date and keep an eye out for registration and programming details on the SMB meeting’s website: 2023.smb.org. 

Mathematical Neuroscience Subgroup

The terms for all current officers of the MathNeuro subgroup will end in July 2023. The subgroup is looking for fresh new people with new ideas to consider running. Note that in total there are 3 to 5 officer positions, including: Chair, Vice-Chair, Advisory Committee Members (at least 1 and up to 3). For further details email Cheng Ly: cly@vcu.edu .


Royal Society Publishing

Journal of the Royal Society Interface publishes high-quality Research Articles and Reviews at the boundary of the life and physical sciences, including epidemiological studies. Our papers apply innovative modelling techniques to address some of the most pressing topics in infectious disease research.

Notable published work includes:

To find out more or to submit your paper, take a look at royalsocietypublishing.org/journal/rsif or email interface@royalsociety.org.


Upcoming Conferences and Workshops

The SMB Annual Meeting is coming up from July 16-21, 2023, at Ohio State University, Columbus, OH, USA.  Submissions for mini-symposia proposals are now open: 2023.smb.org/abstract-submission/. The deadline to submit is March 6, 2023.

Back to the top


By Sara LooImage for People section

Interview with Laura Kubatko, Professor of Statistics and Evolution, Ecology and Organismal Biology at Ohio State University. She is the chair for the upcoming SMB 2023 meeting in July.

Back to the top

Editorial Section

Image for Editorial Section

Producing a good outreach talk

By Thomas Woolley

Before we get into the nitty gritty of how to produce a good outreach talk let me first supply a bit of information about my bona fides. I worked as the Fellow of Modern Mathematics at the London Science Museum, where I had to help redevelop their mathematics gallery. I have presented popular mathematics talks around the world, for example I’ve spoken at MoMath in New York, COSI in Columbus Ohio, Karta Initiative in Mumbai, Dulwich Colleges across China and I opened the Second ever Athens’s Science Festival (a personal highpoint!). Further, I was one of the maths advisor on “Dara O’Briain’s School of Hard Sums”. I have spoken in front of all age groups and even to inmates at Cardiff’s prison.

So, you’d think I’d know what I’m talking about, wouldn’t you? But that’s the thing about outreach: different audiences react differently to the exact material. So the same anecdote has made the city centre of Oxford roar with laughter and died in front of a group of teenagers.

Thus, the following thoughts are by no means “correct”. They are simply musings that I’ve put together through my presenter journey. If they don’t work for you that doesn’t mean that you’re doing it wrong, no more than the list means I’m doing it right, but my hope is that the following ideas will provide you with a set of thoughts on which to build your own presentations.

  1. Be enthusiastic. I cannot stress this enough. Your enthusiasm for a subject matter is contagious and helps to inspire your audience. Show your passion and help the audience see why you love maths.

I once gave a career talk at Techniquest, Cardiff. I was after two engineers: one built tanks, one built formula 1 cars. Both incredible subjects. Yet, it was I who had the biggest crowd of people wanting to talk to me at the end because the engineers bored the audience to tears. At no point was there any sign of passion, just a dry presentation of information about their daily jobs.

  1. Know your audience. It is essential to understand the age group and level of mathematical knowledge of the audience you will be speaking to. This will help you choose appropriate topics and examples that will be both interesting and accessible to them.

I have messed this one up many times. I was once asked to present my “maths of zombie infections” at Coventry University. I thought it would be part of an outreach day. Turned out I was giving it to the Applied Mathematics Research Group. The maths was incredibly trivial for them! Although they still had a fun time, I don’t think it was what they were expecting.

  1. Make it interactive. To keep your audience engaged, consider incorporating interactive elements into your talk. For example, you can use interactive visual aids, quizzes, or group activities to encourage participation and make the material more engaging.

You don’t have to go full style over substance, but the average person’s attention span starts to wain after 15 mins. So think about changing it up every so often. Even just throwing a question out to your audience is a good way of checking whether they’re following your ideas.

  1. Show your relevance. Highlighting real-world applications of mathematics helps an audience understand the importance and relevance of the subject.

Now, you’re all mathematical biologists reading this, so this should be pretty simple. But it’s also something to keep in mind when giving academic talks. Starting from the application should be a no brainer as it will be hopefully something everyone can agree is important, but it is surprising how many talks I see where there is no initial motivation. And if you’re not motivating your audience they’ll just drift away.

  1. Keep it simple.

I always try give my presentations to my nonmathematician wife. She’s certainly my hardest critic and if she can’t understand it then I’ve pitched it wrong.

In conclusion, creating a successful outreach talk on mathematics is more art than science that requires a delicate balance of creativity and expertise. By highlighting the relevance and beauty of maths, you can turn a dry subject into an exciting and enlightening experience for your audience. Remember, the key to a great talk is not just in the information you share but in how you share it. So, go ahead and give a talk that will inspire a generation.

Back to the top

Featured Figures

By Alys Clark

Early Career Feature

Our featured ECR this issue is Alex Browning (University of Oxford) who has recently published a paper titled ‘Geometric analysis enables biological insight from complex non-identifiable models using simple surrogates’ (linked here). We asked Alex to share more about his paper here:

The paper focuses on a key challenge at this model-data interface, which  is to reconcile the complexity and identifiability of models with the information available from data. Indeed, in some cases—clinical data of tumour volumes, for example—issues of identifiability cannot simply be resolved by increasing the quantity or granularity of data. Yet, mathematical models appear poised to play increasingly important roles not just in basic science, but in the clinic by prospectively predicting patient outcomes and guiding treatment decisions.

Model complexity is often a practical constraint: a mathematical model designed to probe the dependence of tumour growth on the spatial availability of oxygen need describe both oxygen and space. However, it is often only simpler models, such as the logistic and Gompertz growth models, that are identifiable from standard experimental measurements. Particularly in the context of tumour modelling, such models are largely phenomenological and contain parameters that relate indirectly to the underlying biology through features observed in data (i.e., the early-time exponential growth rate; the maximum tumour size). Such parameters contrast with those in constructionist models that contain potentially non-identifiable parameters of direct biological interest (i.e., the oxygen consumption rate and diffusivity). Should the simpler model provide an adequate fit, traditional model selection metrics will favour simplicity; at odds with a modelling goal of probing a specific biological mechanism.

To draw insights from complex, non-identifiable models that incorporate key biological mechanisms of interest, we study the geometry of a map in parameter space from the complex model to a simple, identifiable, surrogate model. In effect, the identifiable surrogate models provide a layer of interpretation between the set of non-identifiable parameters and the goodness-of-fit metric traditionally studied in model selection or identifiability analysis. We demonstrate our approach by analysing mathematical models of avascular tumour growth, an area where experimental or clinical measurements are often necessarily limited and insufficient to parameterise event relatively simple constructionist models. Our approach is able to classify non-identifiable parameter spaces into identifiable parameter combinations that relate to phenomenological features, and importantly draw on complex models to provide biological insight in the face of non-identifiability.

Most downloaded article in Bulletin of Mathematical Biology in November 2022

The most downloaded article from the Bulletin of Mathematical Biology in November 2022 was authored by Renee Brady and Heiko Enderling, and was titled “Mathematical Models of Cancer: When to predict Novel Therapies, and when not to”.

This paper looks at trends in cancer modelling research, a field where clinical translation and evaluation of model guided treatment strategies has taken off in the last decade. A key contribution of the paper was to consider appropriate complexity in mathematical models, and to determine when a model is adequately parameterised to predict therapeutic outcomes and when it is not. The authors proposed a pipeline to establish model capability in predicting novel and/or optimal therapies, which is shown in the figure below. Importantly they highlight “shortcuts” that should be avoided when developing translational models. This includes use of mechanistic models that cannot be translated, calibrated, and validated with high-quality and independent datasets.

You can read more about this proposed pipeline in the article, linked here.

Back to the top

Comments are closed.