When I’m not studying Cosmology, I work with UK healthcare scientists to try and improve the treatment of cancers for NHS patients. Whilst you may be surprised to hear of a collaboration between cosmologists and scientists studying cancer, it turns out that we have more in common than you might think.
Multiscale Complexity
One of the most profound phenomena exhibited by nature is multiscale complexity, where complex behaviours emerge from the interactions between physical phenomena at different scales of organisation.
The emergence of multiscale complexity in Cosmology and Cancer is illustrated above.
In Cosmology, I study gravity and dark matter on scales of the entire observable Universe. Analysis is performed on individual galaxies, including prescriptions on how their stars formed. Therefore, this links perhaps the largest scales one can comprehend – the entire Universe – to scales of the solar system we reside in.
In cancer research, the aim is to understand how measurable signals at the genetic level influence healthcare outcomes on a national scale. This involves linking data from a patient’s genome to broader metrics such as proteins and cells. By connecting these diverse measurements, we can gain insights into the complex interplay between genetics and overall health outcomes in the context of cancer care.
Statistics For Multiscale Analysis
For this purpose, I develop statistical methods for multi-scale analysis, which are visually illustrated in the graphic above.
These statistics enable us to analyze small-scale data, like galaxy images or genetic sequences, while connecting them to broader signals of interest, such as cosmological phenomena or healthcare outcomes in clinical trials.
The statistical techniques are implemented in the open source software PyAutoFit, which I developed in collaboration with the Biotech company Concr.Therefore, the same software package is being used to understand the mysteries of the Cosmos and better treat cancer here on Earth.
PyAutoFit is a probabilistic programming language — meaning that makes it straight forward to fit probabilistic models to data, using modern Bayesian inference techniques.
Carcinoma of Unknown Primary Site
My research has a focus on Carcinoma of Unknown Primary Site, or CUP for short.
CUP is where malignant cancer cells are found in a patient but the original cancer is unknown. This makes it extremely challenging to treat, with there currently no approved therapies or immunotherapies in the UK.
I am part of Innovate UK clinical trial on CUP, in collaboration with the Biotech company Concr, the The Christie NHS Foundation Trust and the pharmaceutical company Roche.
The aim is to use multiscale models of cancer therapy with clinical trial data on CUP, in order to develop a predictive model of the otherwise unknown cancer and provide an evidence-based pathway for treatment.
Press:
I have been featured in The Sunday Times and interviewed on live TV about the collaborative research between Astronomers and Healthcare Professionals:
You can watch the full interview at the following link:
https://www.facebook.com/watch/?v=3504354832998051
PyAutoFit Links:
GitHub: https://github.com/rhayes777/PyAutoFit
readthedocs: https://pyautofit.readthedocs.io/en/latest/
JOSS Paper: https://joss.theoj.org/papers/10.21105/joss.02550
Binder (This link runs PyAutoFit in your web browser with no installation, checkout ‘introduction.ipynb’ on the home page): https://mybinder.org/v2/gh/Jammy2211/autofit_workspace/HEAD