The project develops and applies statistical and machine learning techniques for interpreting DNA profiles of plants.
The goal is to build a link between measured intensities of molecular probes hybridised to microarrays and phenotypes, such as crop yield or resistance to drought. The project is developed in collaboration with Diversity Arrays Technology Pty Ltd in Canberra.
What will this research achieve?
The aim is to develop economically viable techniques for improved breeding (crossing) of crops. The project combines advanced machine learning techniques with novel, inexpensive molecular profiling technologies.
- Marker Assisted Selection of Crops for Breeding eg barley, wheat, sorghum, rice, chickpea and sugar cane.
- Potentially huge economical and ecological benefits for agriculturally focused Australian economy.
- The results are applicable to international agriculture, especially in the lesser developed countries.
What are the key features?
Application of machine learning and pattern recognition techniques prediction of plant traits from genetic profiles. This will in particular include application of techniques used in cancer genomics to the field of plant genetics field.
Research objectives
- Development of practical methods/algorithms for transforming of raw genomic profiles of plants into useful knowledge.
- Development of software tools facilitating such a transformation.
- Demonstration of practicality of proposed solutions and methodology via experimental validation (test planting and evaluation).
- Demonstration of feasibility of commercial implementation of proposed solutions.
- Development of word class expertise in plant bio-technology.
Research team
Adam Kowalczyk
Justin Bedo, PhD student