Australasian Applied Statistics Conference 2016

28 November – 2 December 2016, The Windsong Pavilion, Four Winds, Barragga Bay NSW


Invited Speakers

Graham Hepworth

G Hepworth PhotoStatistical Consulting: We may think we are doing well, but how do our clients see it?

Graham Hepworth has worked for over 30 years as a statistical consultant, and is currently Deputy Director of the Statistical Consulting Centre at The University of Melbourne. He began in forestry, then moved to primary industries, where most of his work was in the design and analysis of experiments, and also in sampling. In the university environment the consulting has been broader, covering government agencies and businesses as well as academic research.

Graham’s methodological research interests are mainly in the estimation of proportions by group testing (pooled testing), and in confidence intervals for discrete data. This interest arose initially from consulting in the field of plant virology. His collaborative research, much of which has also arisen from consulting, has included work in animal reproduction, plant pathology, cardiac electrophysiology, dentistry, nursing, compost science, ecology, entomology and neurology. He is also interested in researching statistical consulting itself, and how effective it is for clients.

Graham has taught undergraduate and postgraduate subjects at the university, but in recent years he has concentrated on two intensive short courses for participants from both outside and within the University – one on experiments and the other on surveys. He recently conducted a course for Defence Science and Technology researchers on the design and analysis of simulation experiments.

Karen Meyer

Making the most of precious data in estimating quantitative genetic parameters via restricted maximum likelihood

Karin is a professor at the University of New England, in the Animal Genetics and Breeding Unit, involved in research on quantitative genetics of beef cattle. She holds an undergraduate degree in agricultural sciences (Goettingen), and postgraduate degrees (MSc, PhD and DSc) from Edinburgh university. 
She became interested the use of REML to estimate genetic parameters for livestock early on and has pursued this interest throughout her career. This included model development and applied analyses of large and diverse data sets. Her major impact has been through writing and freely distributing software for mixed model analyses via REML, initially DFREML now superseded by WOMBAT. Karin has authored numerous scientific papers, with a H-index on Google Scholar of 48.

Jay Ver Hoef

JayPhotoSpatio-temporal models for stream networks.

Jay Ver Hoef develops and evaluates new statistical methods for ecological problems, especially those involving spatial and spatio-temporal statistics (e.g. stream data, animal abundance). One of his important methodological achievements was pioneering the development of spatial statistical models for data collected along stream networks. 
|Jay works as a senior statistician for the National Marine Mammal Laboratory, a government research lab within NOAA, based in Seattle. He has adjunct positions at Oregon State University and the University of Alaska, Fairbanks.

He is a Fellow of the American Statistical Association, and has been honoured with a number of awards, including a Distinguished Achievement Award from the American Statistical Association (Statistics in the Environment Section).

Scott Foster

Models for fishy data: two applications of mixture models in fisheries science.

Scott Fisher PhotoScott Foster is a Senior Research Scientist at Data61, a Division of The CSIRO. 
He studied mathematics at The University of Tasmania and went on to study statistics at the University of Newcastle before “getting a real job.”  He did get a real job, well, one that paid him at any rate.
Scott started out as a consulting biometrician with the Queensland Government’s Department of Primary Industries in the far north town of Mareeba. There he was exposed to a wide variety of problems, from agronomy, horticulture, aquaculture and fisheries. 
He liked learning about statistics andits applications.  In fact, Scott liked it so much that he pursued a PhD at the University of Adelaide looking at a statistical problem for beef cattle genetics (how to make them grow faster on less food). 

After finishing he moved back home to Hobart, where he is still.  Scott now works on applied problems of national importance, from marine ecology, biodiversity, fisheries and climate.  The work covers all aspects of statistical application: design, analysis, and communication.  He feels very lucky to be able to work on good (and important) problems, with great scientists, in his home town.

Adam Norman

A Norman PhotoAdam Norman is currently completing a PhD with the University of Adelaide in a joint project with Australian Grain Technologies, Australia’s leading plant breeding company. His project is focusing on the application of genomic selection in a plant breeding program, and ties traditional plant breeding and genetics concepts together with molecular genetics and statistics.
He completed a Bachelor of Agricultural Sciences at the University of Adelaide in 2012, and undertook an Honours project in plant breeding and molecular genetic in 2013

Julian Taylor
Julian Taylor PhotoDr. Julian completed his undergraduate and postgraduate degrees from the University of Adelaide with a focus on statistical and computational modelling of plant breeding experiments. He then went on to complete a postdoctoral fellowship in CSIRO researching variable selection methods for high dimensional gene-trait association analysis. During this fellowship he collaborated closely with Ari Verbyla and they released the R package wgaim, providing users with flexible and efficient QTL analysis software for analysis if complex plant based experiments.

In his current employment, Julian is a research biometrician located in the Biometry Hub of the University of Adelaide and is also Research and Collaborative Projects division leader of the GRDC funded statistical research group, Statistics for the Australian Grains Industry (SAGI). He collaborates and publishes widely with plant research staff within the University of Adelaide and the Australian Centre for Plant Functional Genomics, focussing on statistical genetics projects that include linkage map construction and high dimensional gene-trait association.
Recently, in close collaboration with Dave Butler, he co-authored and released the linkage map construction R package ASMap that uses the MSTmap algorithm for efficient construction of high dimensional genetic marker sets. The package has been widely adopted both locally and internationally providing plant researchers with rapid turnaround time in the construction and diagnosis of linkage maps.

Simon Wood

Simon Woods PhotoSimon Wood works as a Professor of Statistical Science at the University of Bristol, and is author of the R recommended package ‘mgcv’ for smooth additive modelling.
Having trained in physics he worked in mathematical biology, before moving into statistics via statistical ecology. His current research interests are in the mathematical and computational statistics of smooth additive modelling, particularly for models beyond the exponential family, for data with short range auto-correlation, and for very large data sets requiring large models.

Greg Rebetzke

Greg Rebetze PhotoThe challenge of prioritising across phenotypically complex traits in breeding for highly variable environments.

Greg has a background in statistical genetics having graduate studies in statistics and quantitative genetics at Queensland (Masters) and North Carolina State (PhD) Universities.
His research focus has been in delivering genetic understanding in phenotypically complex traits and underpinning the physiological drivers of adaption and performance in rainfed wheat cropping systems. This delivery has included identification of genomic regions and improved breeding/selection methodologies, and germplasm to breeding programs including advanced parental lines and commercial varieties.

Specifically, this work has focussed on genetic improvement of drought tolerance and particularly improved water-use (WUE), and more recently improved nitrogen-use efficiency (NUE). The high WUE trait is an important target as water-limited performance is improved without cost to productivity in favourable seasons.
Dr. Rebetzke has demonstrated that differing traits contribute to greater WUE owing to changes in the timing and availability of water with location and season across the Australia wheatbelt. Similarly, the genes underpinning these traits are not the same and are complex to screen phenotypically thereby slowing their adoption in commercial breeding.

Nicole Cocks

Nicole Cocks PhotoThe challenge of prioritising across phenotypically complex traits in breeding for highly variable environments.

Nicole completed her undergraduate studies in medical mathematics and then went on to achieve First Class Honours with a major in statistics at the University of Wollongong.
Prof. Brian Cullis and Dr. Emi Tanaka co-supervised Nicole’s honours thesis for which she aimed to improve the computational efficiency of model based design with an application to multiphase experiments following her success in being awarded a UHS by the GRDC.
In her new role at CBB as an associate research fellow, Nicole is responsible for collaboration with other agricultural scientists and is funded by the GRDC.

She is currently involved in the analysis of survey data collected on Pacific Island sea cucumbers fishers and the GRDC funded MEF project that has been in operation since 2010.