Using pathology data in a smarter way

Tony Badrick of Bond University gave this seminar at the John Curtin School of Medical Research on 23 May. His talk included some of his current projects with Brett Lidbury and myself, as well as others looking at reference intervals and other issues around the mass of data generated by pathology laboratories.

Grant proposal boot camp, University of the Sunshine Coast

I could swear the the same road works on the freeway from Brisbane are still under way, many months since my last visit to USC. The purpose of this May trip was to whip an OLT grant proposal into shape. Five people, eleven hours, and multiple cups of coffee later, it’s looking pretty good. Due for internal review at the end of the month.

Estimating the economic impact of a free trade agreement on ASEAN. Case study: Vietnam and the ASEAN-Australia-New Zealand free trade agreement.

Tai Pham is a PhD student in the Faculty, and I’m on his supervisory panel. Tai gave his initial seminar on the 9th of May, and he seems to be well on top of the input-output model he’s going to use, and the software that goes with it. Questions centred around the possibilities of sensitivity analysis around the nuances of the proposed model, and making sure that the project carves out its unique contribution in a clear manner.

Rational thoughts: an invitation to number theory, geometry and computation

Adjunct Professor Stephen Glasby from UWA was visiting UC 5-9 May, and he gave this talk on 9 May of broad interest to mathematics students, mathematics educators and mathematicians alike. We were encouraged to think deeply about simple things, namely the rational numbers.

Big data in genomics: implications for reproducible computation

A/Prof Gavin Huttley of the John Vurtin School of Medical Research at ANU came to UC on 9 May to help celebrate Big Data Week. His talk took us through the growth of big data (big has got bigger in the last twenty years!) then to the principles of reproducible computation. Some of the horror stories about papers withdrawn because of software errors that Gavin recounted certainly helps focus our minds on the importance of the issues he raised. New software products such as Python could certainly help – maybe I will have to tackle learning this package quite soon.

OSP report: statistical contributions to clinical chemistry

My turn to give a talk, on 2 May, to the faculty, closing the loop on my study leave. I spoke about the Alzheimer’s, CFS and hepatitis papers, as well as the less developed work on liver function tests. There were some digressions too, about everything from personal-omits to the mechanical Turk! Questions at the end centred around the liver toxicity data mostly – maybe I didn’t explain it very well. Also Roland Goecke was interested in the possibilities of adding value to the CFS cohort by using his facial expression software.

Model selection in mixed linear models

Alan Welsh of ANU gave this talk to the Statistics Society canberra Branch meeting on 29 April. Though he restricted his discussion to the simplest of models, the Laird-Ware model with one level of grouping, there were still plenty of issues to chew over. Whilst my Biometry students might be disappointed at the multiple directions and multiple evaluation methods available, I really appreciated Alan’s measured delivery and clear exposition. There’s a paper in Statistical Science too.
Discussion at dinner wandered on to the statistical inference reading unit I teach, and alternative choices of textbook (Hogg and Craig a leading contender).