Developing Interpretable Machine Learning Models For Clinical Imaging And Single-cell Genomics
Funder
National Health and Medical Research Council
Funding Amount
$1,312,250.00
Summary
Machine learning methods will be vital to make best use of the deluge of data generated by high-throughput technologies in biomedical science. To get the most out of these models, however, we need to be able to unpack the 'black box'. I will use curated clinical and public research data to benchmark and develop interpretable deep learning models and software tools. These models will be used for breast cancer screening programs and for analysis of complex, large-scale single-cell genomics data.
Preparing Australia For Genomic Medicine: A Proposal By The Australian Genomics Health Alliance
Funder
National Health and Medical Research Council
Funding Amount
$25,000,000.00
Summary
The sequencing of the human genome brings the possibility of more accurate identification of the underlying basis of many diseases. This technology has moved so rapidly, however, that clinical access has been limited. In this application, a national alliance of clinicians, researchers, health economists and policymakers will evaluate the case for clinical genomics across inherited disease and cancer, determine how best to deliver this to the patient and train a capable workforce.
Transcription factor – enhancer – promoter based regulatory networks. This project aims to develop new understanding on how multicellular organisms (including humans) develop, and how mutations in distant regions of the genome can affect human traits. The way the human genome is interpreted by the cellular machinery is still a mystery. We have a reference sequence and know where the majority of coding genes are, but we are far from understanding how the genome is regulated to generate the divers ....Transcription factor – enhancer – promoter based regulatory networks. This project aims to develop new understanding on how multicellular organisms (including humans) develop, and how mutations in distant regions of the genome can affect human traits. The way the human genome is interpreted by the cellular machinery is still a mystery. We have a reference sequence and know where the majority of coding genes are, but we are far from understanding how the genome is regulated to generate the diversity of cell types in our bodies. Enhancer regions interact with proximal promoters to regulate gene expression level and tissue-specificity. This project aims to develop transcriptional regulatory network models using high throughput chromatin interaction data and expression perturbation to link promoter and enhancers genome-wide.Read moreRead less
Automation of metric temporal reasoning. A major contemporary engineering concern is to ensure the predictable and robust operation of computer systems involving software, hardware, and human users. The need for systematic and careful construction of such systems requires the development of formal methods based on a dense view of time rather than the traditional step-by-step models.
automated strategic reasoning. Formal methods are used to ensure robust correct behaviour in design and implementation of computer systems. Traditional models of computer operation involve a linear sequence of behaviour but today’s systems are complex interactions between many components including the environment of the system and human users. Thus analysis is done via a logical game between components where each is trying to meet its specified requirements regardless of what others do: formalis ....automated strategic reasoning. Formal methods are used to ensure robust correct behaviour in design and implementation of computer systems. Traditional models of computer operation involve a linear sequence of behaviour but today’s systems are complex interactions between many components including the environment of the system and human users. Thus analysis is done via a logical game between components where each is trying to meet its specified requirements regardless of what others do: formalisms include branching time and competing coalitions of agents. This project is to take early advantage of recent breakthroughs in automated logical reasoning with such models by the investigator to deliver general practical techniques of system development and verification.Read moreRead less
Next generation high throughput lipidomics using adaptive modelling. This project aims to develop a unique high-throughput method to capture the lipidomic profile of human plasma suitable for large human population screening. Lipids are fundamental to every biological system, but our understanding of their regulation in humans have been largely superficial. By incorporating a new lipidomics approach, with genomic data, this project aims to expand our understanding of human biology by identifying ....Next generation high throughput lipidomics using adaptive modelling. This project aims to develop a unique high-throughput method to capture the lipidomic profile of human plasma suitable for large human population screening. Lipids are fundamental to every biological system, but our understanding of their regulation in humans have been largely superficial. By incorporating a new lipidomics approach, with genomic data, this project aims to expand our understanding of human biology by identifying regulators of lipid metabolism. The large diversity in humans necessitate sufficient sample sizes to identify true genetic regulators, but to date techniques capturing phenotypic data (lipids) have been largely limited. It is anticipated that this study will identify new regulators of lipid metabolism in humans.Read moreRead less
Who’s who in the plant gene world? As many more plant genomes are sequenced, the bottleneck is being able to interrogate and translate this data into applications for crop improvement. This project will develop and apply a population graph database, hosting genome data for the world’s major crops and their wild relatives, allowing the characterisation of gene diversity on an unparalleled scale. Analysis of this data will reveal the presence/absence and sequence diversity for classes of genes for ....Who’s who in the plant gene world? As many more plant genomes are sequenced, the bottleneck is being able to interrogate and translate this data into applications for crop improvement. This project will develop and apply a population graph database, hosting genome data for the world’s major crops and their wild relatives, allowing the characterisation of gene diversity on an unparalleled scale. Analysis of this data will reveal the presence/absence and sequence diversity for classes of genes for important agronomic traits including disease resistance, flowering time and legume nitrogen fixation which will enable plant breeders to identify and apply novel genes and allelic variants for use in breeding programmes, accelerating the production of improved crop varieties.Read moreRead less
Symmetry and computation. The overall objective of the project is to explore connections between symmetry and computation, especially the theory and algorithms that facilitate the use of groups in computational science. The main outcome will be theoretically fast algorithms and implementations to drive applications in the sciences and for secure communication.
Accelerating pulse breeding using machine learning. Advances in genomics and high throughput phenotyping are generating vast quantities of data that can be applied for crop improvement, however the lack of computational analysis tools and approaches limits the full exploitation of this data. Pulse legumes are currently under utilised in Australian agriculture due to poor adaptation, however they offer significant benefits both for soil improvement and the production of high protein crops. This p ....Accelerating pulse breeding using machine learning. Advances in genomics and high throughput phenotyping are generating vast quantities of data that can be applied for crop improvement, however the lack of computational analysis tools and approaches limits the full exploitation of this data. Pulse legumes are currently under utilised in Australian agriculture due to poor adaptation, however they offer significant benefits both for soil improvement and the production of high protein crops. This project will develop machine learning (ML) tools for the analysis of pulse legume crop traits and their association with genomic variation to accelerate the breeding of high performance pulse legumes for Australian growers.Read moreRead less
IDENTIFYING CONTROL ELEMENTS IN CHLOROPLAST GENE EXPRESSION. Energy from sunlight is captured by photosynthesis in plants, providing the basis for the terrestrial food chain. This process takes place in chloroplasts, subcellular structures that derived from photosynthetic bacteria a billion years ago. Chloroplasts have their own DNA, containing genes encoding the most important photosynthetic proteins. This project aims to provide the world’s best resources for the study of chloroplast genes. In ....IDENTIFYING CONTROL ELEMENTS IN CHLOROPLAST GENE EXPRESSION. Energy from sunlight is captured by photosynthesis in plants, providing the basis for the terrestrial food chain. This process takes place in chloroplasts, subcellular structures that derived from photosynthetic bacteria a billion years ago. Chloroplasts have their own DNA, containing genes encoding the most important photosynthetic proteins. This project aims to provide the world’s best resources for the study of chloroplast genes. In the process, we will discover how these important genes are regulated to provide photosynthetic proteins in the right amounts, in the right cells, at the right time. The knowledge and resources gained will facilitate improvement of photosynthetic function in future agricultural crops.Read moreRead less