Postgraduate profiles: Full alphabetical listing


Our postgraduates carry out interesting and often vital research into all manner of subjects across all research areas. Some are interested in politics, others in law, epidemiology, seagrasses or information and communications technology.

To promote their efforts, and to encourage others who are inspired to make their own mark on the world, we present the work of our current and past postgraduate students here.


Robyn Anderson


Start date

Mar 2019

Submission date

Robyn Anderson

Robyn Anderson profile photo


Using Deep Learning for Trait Prediction in Brassica napus (Canola)


Plant traits are the result of genetic and environmental interactions, and understanding the basis of these traits is an important step to improve crop breeding. Finding the basis for desirable traits (such as yield, drought resistance and disease resistance) in plants is difficult as the processes that underlie these traits are complex and not fully understood.

Methods such as Genome Wide Association Studies, Marker Assisted Selection and Genomic Selection are being used to inform the breeding process by finding associations between genetic markers and traits of interest, however progress has been slow, and these methods are not capable of capturing all of the underlying influences, which further slows the production of new crop varieties with improved traits.

Deep Learning is a field within Artificial Intelligence that has the capacity to 'learn' data and find the mathematical patterns and relationships within the data. Brassica species include several agronomically important plants, consequently relatively large genomic and phenotypic data resources are available, which are required to build accurate deep learning algorithms. My project will involve applying deep learning methods to crop breeding data from the Brassica napus (Canola), to determine if deep learning can help us understand the complexity of plant traits and improve crop breeding programs outcomes.

Why my research is important

The predicted effects of climate change on agriculture are devastating, and our ability to feed the growing global population relies on many advances within agricultural practices. A major contributor to this is crop improvement, which has the ability to create varieties that are more drought, stress and pathogen resistant which will help maintain food security.

The use of deep learning methods on genetic data is still new, and the best methods to apply are not yet well explored. The recent literature applying deep learning to this kind of data has involved a variety of methodology and little standardisation. I intend to explore the use of these different methodologies and evaluate the differences between these approaches within my project.