Machine Learning

Metformin use is associated with lower mortality from bacterial sepsis and improved immunocompetence in Thai diabetes patients with acute melioidosis
Metformin use is associated with lower mortality from bacterial sepsis and improved immunocompetence in Thai diabetes patients with acute melioidosis

Lay summary Melioidosis is a neglected tropical disease caused by the bacterium Burkholderia pseudomallei, which can be acquired from the environment through contact with soil or water. Type 2 diabetes is a major risk factor in acquiring the infection, though some past studies showed that some diabetes medications prevent deaths from melioidosis. However, although metformin is the most common treatment for diabetes, we don’t know whether it specifically prevents deaths from melioidosis.

15 Nov 2025

PhD Project
PhD Project

My PhD project was conducted under the joint supervision of Dr Diego Oyarzún (computational) and Prof Peter Swain (laboratory & computational), at the University of Edinburgh. My project examined the metabolic cycle in budding yeast. This yeast metabolic cycle is defined by levels of metabolites in the cell increasing and decreasing in regular cycles, which are linked to cell division. Changing feeding conditions for the yeast cells or removing genes from the yeast genome that are responsible for metabolic processes affect the speed of these metabolic cycles.

31 Oct 2023

Accuracy and data efficiency in deep learning models of protein expression
Accuracy and data efficiency in deep learning models of protein expression

Deep learning, a type of machine learning, can be used to predict how much protein a specific DNA sequence can produce. But, it usually needs a lot of data. We managed to get a deep learning model to do make predictions with less data, especially if the data is diverse. Our study is useful because people can then produce fewer DNA to train a model, and this is cheaper.

15 Dec 2022

aliby
aliby

aliby is a Python-based end-to-end yeast microfluidics image analysis pipeline developed mostly by Dr Alán F. Muñoz during his PhD a the University of Edinburgh. I wrote the metadata parser, helped test previous versions of the software (and found bugs), and trained research group members on how to use it.

13 Dec 2021