UW00005 - Statistics for the Australian Grains Industry (II)

This project will deliver statistical support and training and develop innovative statistical technologies and statistical software to the grains industry of Australia. Its primary remit is to provide high-level support and training to national, public pre-breeding and breeding programs and the National Variety Testing (NVT) system, Australian Wheat and Barley Molecular Marker Program, Molecular Markers for Pulse Breeding Program, Canola Molecular Marker Program and other GRDC-funded research projects.

The project will provide statistical technical expertise to the Australian grains industry through three divisions: an industry liaison and training (ILT) division, a research and collaborative projects (RCP) division and a software development and implementation (SDI) division.

The ILT Division will be responsible for the management of statistical resources, statistical training and capacity building across a broad range of GRDC-funded projects. There will be two units within this division. The Industry Liaison Unit is largely concerned with quality control. It will identify statistical, software and data management requirements for GRDC-funded projects to enable appropriate allocation of resources within the project. The Education and Training Unit is concerned with the development and extension of resource material in order to build or strengthen within-project capacity in the area of agricultural statistics. Both units will utilise the new website MM-onTAP (Mixed Models-Theory and Practice) that has recently been developed by Statistics for the Australian Grains Industry (SAGI) staff.

The website aims to provide information about mixed models in terms of theory and practice, with particular application to plant and animal breeding data. It will be used extensively by the Training Unit as a repository for resource material and for advertising up-coming events such as courses, seminars and other information and training sessions. The website includes a system for managing the SAGI2 report series, so will be a crucial tool for the quality-assurance component of the Industry Liaison Unit. The ILT Division will strengthen links between SAGI2 and current collaborators and create new alliances with a wider range of GRDC-funded research projects. It is expected to lead to more effective collaboration, a greater understanding of the importance of statistics and a greater level of statistical expertise across a broader section of the research community.

The end result will be more rigorous, thence cost-effective, GRDC-funded research in the short term and well into the future.

The RCP Division comprises two units, the Research Unit and the Collaborative Projects Unit. The remit of the Research Unit is to increase the effectiveness of GRDC-funded research projects through the development of new statistical methodologies with application to agricultural and biological research. One of the additional strengths of the research unit in SAGI2 will be the strengthened linkages and collaborations with international researchers working in related areas. In particular, Professor Fred van Eeuwijk from the Wageningen University will present two courses in QTL analysis in July 2011. Whilst in Australia, Professor van Eeuwijk will scope collaborative work with staff of this unit and staff of the SDI Division. This work will involve the development of new tools for genome-wide selection and validation, as well as migration of existing map construction and QTL identification software from GENSTAT to R.

We will continue to build on our relationships with other international statisticians. In particular, we will work with Professor Robin Thompson, Rothamsted Research and Dr Christine Hackett at the Scottish Crop Research Institute. Collaboration with Professor Thompson will be in whole-genome analyses and related mixed-model implementations. Collaboration with Dr Hackett will continue extensions of techniques for identification of QTLs in multi-treatment situations. The Research Unit will also undertake work improving the modelling of G x E for local area prediction in the National Variety Trials (NVT) system and advanced approaches for optimal experiment design.

The SDI Division is responsible for the implementation of new statistical tools into user-friendly software. Major activities will include the development and release of a package for generating experimental designs of use in agricultural and plant genomic research, ongoing improvements to the mixed-models software program, ASReml, and migration of genomic analysis tools from GENSTAT to R. There are two units in this division and both link across the project to capture and package advances in statistical methodology, support industry liaison and training initiatives and enhance data integrity through a data-management framework.