Since 2004 I have directed the Biomedical Data Science Laboratory in the UTS Australian Artificial Intelligence Institute. The central question for my lab is how software can improve knowledge discovery and decision-making in large, distributed open systems, especially in medicine and biology.
We are interested in data analytics. Data analytics is the analysis of large databases to find novel, commercially valuable and exploitable patterns. It has been credited with high-profile successes including locating Ivan Milat (the infamous ‘backpacker killer’), predicting the 2012 US election result, and everyday saving companies millions of dollars through fraud detection and improved strategic intelligence. Data analytics is also used in medicine to predict which patients will respond to treatments.
My research is in bioinformatics and data mining that focuses on biomedical datasets. I investigate how to better diagnose and treat paediatric cancer, namely acute lymphoblastic leukaemia and neuroblastoma. Acute lymphoblastic leukaemia is the most common childhood cancer. However, around 20% of the children judged by clinicians to have a mild form and treated accordingly will die. My research aims to help clinicians identify which children are in this category. We aim to create a tool for 'personalised medicine' where therapies are tailored to individuals. We create statistically valid visualisations of patients based on their gene activity patterns, genetic variations, clinical data and treatment outcomes and objectively compare new patients with previous patients treated possibly on different drug trials. Neuroblastoma is an aggressive paediatric cancer with a survival rate of 30% and is the second most frequent cause of cancer mortality in children. Treatment is determined by identifying and counting histological structures in the scanned images of tumour samples. My research semi-automates this process by developing a Computer-Aided Diagnosis system, the first for neuroblastoma, to assist pathologists in their work.
Since 2010 I have been working with parasitologists to develop an in silico vaccine discovery pipeline for several parasites of humans and cattle. Starting from gene sequences in online databases, the pipeline accumulates evidence derived from the bioinformatics tools to identify which protein subunits would make likely vaccine candidates.
We are interested in data analytics. Data analytics is the analysis of large databases to find novel, commercially valuable and exploitable patterns. It has been credited with high-profile successes including locating Ivan Milat (the infamous ‘backpacker killer’), predicting the 2012 US election result, and everyday saving companies millions of dollars through fraud detection and improved strategic intelligence. Data analytics is also used in medicine to predict which patients will respond to treatments.
My research is in bioinformatics and data mining that focuses on biomedical datasets. I investigate how to better diagnose and treat paediatric cancer, namely acute lymphoblastic leukaemia and neuroblastoma. Acute lymphoblastic leukaemia is the most common childhood cancer. However, around 20% of the children judged by clinicians to have a mild form and treated accordingly will die. My research aims to help clinicians identify which children are in this category. We aim to create a tool for 'personalised medicine' where therapies are tailored to individuals. We create statistically valid visualisations of patients based on their gene activity patterns, genetic variations, clinical data and treatment outcomes and objectively compare new patients with previous patients treated possibly on different drug trials. Neuroblastoma is an aggressive paediatric cancer with a survival rate of 30% and is the second most frequent cause of cancer mortality in children. Treatment is determined by identifying and counting histological structures in the scanned images of tumour samples. My research semi-automates this process by developing a Computer-Aided Diagnosis system, the first for neuroblastoma, to assist pathologists in their work.
Since 2010 I have been working with parasitologists to develop an in silico vaccine discovery pipeline for several parasites of humans and cattle. Starting from gene sequences in online databases, the pipeline accumulates evidence derived from the bioinformatics tools to identify which protein subunits would make likely vaccine candidates.
Research Interests
My research interests include:
- Bioinformatics and biomedical data analytics
- Data mining and data analytics
- Machine learning and more broadly AI
- Dimensionality reduction and/ visualisation
- Text analytics
- Social network analysis
- Evolutionary Computation
Links
Here are links to my