McGee Honours Projects

High-throughput phenomics of the vertebral column

Full time or Part time

Background: Machine learning techniques offer great promise for dramatically increasing the size and speed of data collection in the biological sciences, including automated counting and identification from images.

Project Aims: The student will train a machine learning algorithm to quantify the type and number of vertebral elements across a dataset of thousands of images, then examine this data in an evolutionary context using phylogenetics.

Techniques:  Computer skills, including: Python,  Phylogenetics in R

*****

Colour pattern evolution in fishes

Full time or Part time

Background:  Animal coloration has prompted intensive study by biologists over the last several hundred years, but large datasets and evolutionary trees are required to accurately understand what factors promote colour pattern biodiversity.

Project Aims:  The student will compile photographs of marine fishes, then train a machine learning algorithm to quantify colour pattern diversity. They will then analyse colour pattern diversity in a phylogenetic context across a large evolutionary tree of fishes.

Techniques:  Computer skills, including: Python, Phylogenetics in R

*****

Integrating fossils with genomes to infer divergence times

Full time or Part time

Background:  Molecular phylogenetics has dramatically improved our ability to understand evolutionary relationships between groups of organisms, but it is necessary to combine this information with fossil data to accurately estimate divergence times.

Project Aims:  The student will build a phylogeny based on whole genomes, then combine this information with fossil data to accurately infer divergence times across vertebrates.

Techniques: Computer skills, including: R, Phylogenetics software, including RAxML, IQ-TREE, BEAST, and RevBayes

*****

Using computer vision to understand animal movement

Full time or Part time

Background:  Quantifying animal movement and behaviour with video data is often challenging and slow with conventional methods, but new advances in computer vision can dramatically speed up the process.

Project Aims:  The student will utilize a database containing thousands of videos of fish feeding behaviour and develop a means to accurately quantify behaviour with supervised machine learning techniques.

Techniques:  Computer skills, including:  R, Python

*****

Using computer vision to quantify 3D skull data

Full time or Part time

Background:  Three-dimensional imaging techniques allow for detailed measurements of biological structures, but analysing such data is time-consuming. New computer vision techniques offer ways to dramatically speed up data collection.

Project Aims: The student will use pre-existing CT scan data for a range of vertebrates in conjunction with newly available supervised machine learning techniques to automatically measure various skull traits.

Techniques:  Computer skills, including:  Python, R

*****

Identifying invasive species using large biodiversity datasets

Full time or Part time

Background:  Invasive species drive extinction events, habitat destruction, and economic damage worldwide, but there is still much debate on mechanism.

Project Aims:  The student will integrate a large evolutionary tree of fishes with information from biodiversity databases with information on invasive species to create large statistical models predicting invasion risk and potential damages in Australia and worldwide.

Techniques: Computer skills, including:  Statistics and phylogenetics in R

*****