Chad Dechow, an ASReml user from the Department of Dairy and Animal Science at The Pennsylvania State University, studied herd heritability estimates in over 20,000 herds. He and his colleagues compared covariance parameter estimates in subsets of data, pooled from herds with high, medium or low individual herd heritability estimates. The individual herd heritability estimates, generated from a regression model using ASReml, applied to milk yield, fat yield, protein yield and somatic cell score. These regression techniques stratified the herds by heritability; additive genetic variance increased progressively and permanent environmental variance decreased as herd heritability increased. Future work by the same authors demonstrated that herds with low heritability had a high rate of parent misidentification and that such herds were poor candidates for genetic testing herds.

“We used ASReml,” says Chad, “because it is a very flexible mixed models program that is capable of analyzing large datasets in a very efficient manner.”

ASReml was used to generate individual herd heritabilities; the model and results can be seen in the original paper published in The Journal of Dairy Science 2008. 91:1648-1651

It’s important for breeders to know which factors are genetic and which are affected by environment, for example health events have low heritability indicating that improved management would be important in improving animal health, rather than genetic selection. Understanding which factors can be changed by genetic selection and which through environmental influences is vital as it means that breeders and farmers can plan and manage their herds more effectively; as they understand which aspects can be changed through a breeding programme and which through environmental factors. As heritability differs so significantly between different traits it’s again vital to understand which aspects are most likely to be changed through a breeding program; the higher the heritability a trait is, the more quickly it can be changed through genetic selection; again helping breeders and farmers to plan better, as it can indicate timescales for any genetic changes to take place.

There is no doubt that, as with many things in life, the better understood something is the more effective the planning will be. Research, such as Chad’s is vital to help cattle breeders and farmers breed and manage their herds more effectively. And all good research needs support and suitable tools; ASReml is such a tool. Software that was developed by researchers in animal and crop breeding, it is highly suited to this purpose providing users with the ability to obtain reliable predictions of genetic values. Research needs trustworthy and reliable tools and ASReml is a great fit for the job.

Thanks to Chad Dechow for his help on this piece. The full paper is available here.

Also of interest:

Chad is the editor of The Specht Report - a report on breeding performance.

The Western District of Victoria is known for its heavy clay soils, that are prone to short periods of waterlogging for 2-3 months a year in the Winter; but with the right pasture species livestock producers could increase pasture production during the Summer by utilising this stored soil moisture. Additionally a species that can use out of season rainfall and moisture stored in the soil, would reduce groundwater recharge. One such plant is Summer-active tall fescue.

The experiment, which began in November 2004 with the establishment of the grass, is taking place at the DPI EverGraze research site at Hamilton;  4 grazing system treatments were imposed in a randomised design in September 2006. Since then the plots have been grazed at different levels and herbage mass measured monthly to calculate herbage accumulation.

GenStat’s REML technique was used to model the repeated measurements over time.

This enabled seasonal changes in pasture productivity and quality to be detected, and indicated how the plant responds to climatic events, such as the summer rainfall that is a common characteristic of the research site. Stocking rate, and major husbandry practises, such as lambing and calving, could then be timed to coincide with pasture availability. Within each time period, GenStat was used to determine the effects of grazing treatment on pasture persistence, productivity and quality.  Grazing management was then adjusted, based on the results of the GenStat analysis, to achieve maximum consumption of high quality pasture by livestock, whilst also ensuring the long term survival of the pasture.  Throughout this research, efforts were made to reduce error, and increase the accuracy of the results.  This was achieved by using repeated measurements from fixed quadrants, and also by using GenStat’s graphics capabilities to generate residual plots, which were used to check the normality of the data, and, where needed, impose data transformations.

The study found that the accumulation rate and nutritional value of the Summer-active tall fescue are closely related with its growing conditions. Specific grazing systems and the environmental conditions had an effect on the herbage accumulation and nutritional value. Research, such as Margaret’s is vital in being able to provide sound advice and reliable observations to landowners, and because GenStat was originally developed by statisticians working in agricultural research, it is the most suitable data analysis tool for agricultural research. Couple GenStat’s understanding of the agricultural researchers issues with the reliable and proven statistics, and ethics of good statistical practice and you know your analysis results can be trusted.

Details of Margaret’s project and more information is available at EverGraze. An action sheet based on Margaret’s work on summer active Tall Fescue is here.

And yet, thanks to developments in agricultural research, studying all aspects of the problems and issues faced in agricultural production, farmers and agricultural producers are able to plan more effectively and manage their businesses using far more informed decisions than in the past.

In relation to the issues facing potato growers, research conducted by The Nematology and Entomology Research Group at the Crop and Environment Research Centre at the Harper Adams University College in the UK, into the relationship between the potato cyst nematode Globodera rostochiensis (G. rostochiensis) and the diseases caused by Rhizoctonia solani (R. solani) has indicated there are links between the two problems. Both problems can cause severe damage to a potato crop, with deformed or malformed plants that are not marketable as well as producing reduced yields or even potatoes with a lower nutrient value. It’s estimated that potato cyst nematodes could cause annual losses in the region of 300 million Euros to potato producers in the European Community alone; so a better understanding of the problem, which could lead to a way of controlling this threat, could mean huge cost savings and increased production.

The trial, undertaken by Dr. Matthew Back, Placement Manager (Agriculture) / Lecturer in Plant Pathology and the team at Harper Adams University College in 2000 and 2001, was the first to look at the issues under field conditions, rather than in a controlled environment. By looking at the nematodes and the fungus in natural conditions, researchers are able to determine far better how significant the interaction is between the two problems. Two sites were used, and planted with the cultivar Desiree and the plants were monitored for diseases related to the R. solani.

The results were analysed in GenStat. Regression analysis (again in GenStat) showed there were relationships between the density of the nematodes and the incidence and severity of disease caused by R. solani. Likewise simple linear regression analysis showed a significant linear relationship between the tuber yield and the infection of runners with R. solani. Multiple regression analysis showed that interactions between the nematode root invasion and the runner infection had a significant effect on the final tuber yield. The study has shown that there are clear interactions between this nematode and the fungus. There were positive relationships between the nematode densities and infection on all potato parts, but specifically between the invasion of potato roots by juvenile nematodes and the infection of runners with R. solani. It seems that the interaction is indirect given that each problem affects different parts of the potato. Interestingly it seems that the nematode plays an important role in the development of R.solani diseases, but these infections may reduce nematode development.

Studies such as these are vital in planning for future agricultural development. By understanding the relationship between different pests and pathogens, more precise and hopefully more productive growing strategies can be developed.

Full paper published in: European Journal of Plant Pathology, Issue Volume 114, 2 / February, 2006, pages 215-223.

But a bee is not just a bee; different strains and species of bee have different characteristics and behaviour patterns, so bee keepers and breeders need to be able to identify specific strains in order that they breed the most suitable bees for their requirements. There are obvious physical aspects that can be used such as size, colour or shape, but these methods are not foolproof and often hide specific strain differences. Another option is DNA testing, but this is an expensive and time consuming exercise, therefore a method first identified and introduced before the 1960’s can be used, known as morphometry.

Morphometry is a study of bee anatomy to establish race or strain characteristics that are otherwise difficult to ascertain. The types of anatomy studied are the wings, tongue length, Tomentum width (width of the band of hair on the body segments) and hair length. The most common test and study is on the bee wings, looking at vein formation. Specifically bee breeders look at the “cubital index”, which is the ratio between two vein segments of the cubital cell in a bee’s wing, and the “discoidal shift”, which measures the position of the discoidal joint in relation to the perpendicular through the distal lower joint in the radial cell.

The concept of understanding which bee strain or race is important for bee breeding; as the more purer bee strains can have a better temper (this is not always true, as some inbred bees can be very cross), and therefore there is a less chance of being stung, likewise they seem to keep cleaner and more organised hives than their hybrid cousins. Beyond these “housekeeping” issues there is the bees actual performance, for example researchers have found that certain strains of bee respond quicker to Spring stimulation than others, this is important if the bees are being used for pollination, as it means some strains of bees are more active earlier than others.

The more a bee keeper or bee breeder knows about his bees the more sure he can be about how they will perform and behave. And it seems beewing morphometry can give an indication of the strain or subspecies of Apis mellifera.

The Dark European honey bee, or Apis mellifera mellifera, is defined by having a cubital index of no more than 1.9 and the discoidal shift angle of no more than 0 (although these are ideal standards and not always attainable). Results from the measurements from any hive or colony are displayed on a scattergram and bee keepers and breeders can see how pure their bees are. Jacob Kahn has taken this research further and used GenStat to analyse the measurements more in keeping with quantitative population genetics. He doubled the sample size of 30, normally used by honey bee morphometrists, to a sample size of 60 . 10 of these have had their measurements checked for goodness of fit of normal distribution, and the results showed that only one did not conform to normality. He then looked at any correlation between the cubital index and discoidal shift and found that 6 samples displayed correlation. Jacob hopes that further research and work on this will verify his findings, which would suggest that any correlation indicates genetically purer samples. Jacob suspects that there is a misconception, especially among amateurs, as to what is pure and what is a hybrid. “The term hybrid is generally misunderstood by beekeepers,” says Jacob, “We can talk here about two levels of hybridization: a cross between species can be termed hybridization, but a cross between two strains within a subspecies carrying different alleles of particular loci can also be said to be hybrids. and it is this aspect of population genetics which I am hoping to clarify with the help of GenStat.”

At the moment bee keepers and breeders are content with looking at simple scattergrams to check for bee purity, however Jacob’s work shows that there maybe even more details to discover about bee species or strains. Either way GenStat’s sound statistics has enabled Jacob to dig deeper into the data and initial analysis to find out more about the bees. This additional level of information should allow better definition of variants of Apis m. mellifera which in turn gives bee keepers and breeders greater information on their hives and colonies so they can breed better performing bees.

Read more about Jacob and the work of the Galtee Bee Breeding Group.

Darwin was one of the first to propose that the origin of these differences may be explained by sexual selection, such as traits that have evolved through male to male combat. However, the way in which sexual dimorphism actually evolves from a genome that is almost entirely shared between the sexes remains puzzling. That is because male and female traits are controlled by the same set of genes, thus the influence of selection in one sex is expected to be countered by opposing selection in the other sex, a situation termed intralocus sexual conflict.

A classic example of sexual dimorphism, which may have evolved through such male-to-male combat, are the horns of sheep, which are significantly larger in males than in females. The bighorn sheep (Ovis canadensis) is a species found in North America and Siberia, aptly named for their massive, curved horns of males which can weigh up to 15kg (as much as the rest of the bones in the male body). Ewes also have horns but they are much shorter and have less curvature. Prior to the rutting season the rams create a hierarchy for access to the ewes for mating through contests settled by clashing horns. It’s clear that large horns play an important role in determining mating partners for males, however very few studies have been carried out on the considerably smaller horns of the ewes, which seem to have no known fitness benefit. Possibly the horns on ewes exist because of a genetic correlation with male horns, or perhaps they are a defence against predators.

Researchers in Canada, the United States and Scotland chose to look more closely at this phenomenon, and tested for sexual conflict over horn size and body mass in a wild population of bighorn. The study population inhabits Ram Mountain in Alberta, Canada. The 35 year study captured data on rams and ewes including measurements of body mass as well as horn length and circumference at its base to calculate horn volume.

The researchers used an animal model (a form of mixed model incorporating pedigree information, where the phenotype is modelled as a sum of its additive genetic value and other random and fixed effects) and restricted maximum likelihood in ASReml to determine additive genetic (co)variance for male and female traits. ASReml is well suited to this area of work having been developed by statisticians in animal breeding. Its fast and efficient algorithms enable analysis of large and complex data sets such as the one in this study.

The study tested for sexual conflict by estimating quantitative genetic parameters and sex-specific selection coefficients for two sexually dimorphic traits – body mass and horn volume. The traits showed significant additive genetic variance and were positively, but not perfectly, genetically correlated between the sexes. There was also no evidence of sexually antagonistic selection. The intralocus sexual conflict has therefore been resolved, as sex-biased genetic variation has allowed sexual dimorphism to evolve to a level that is currently somewhat satisfactory to both sexes.

The full paper was published in The Proceedings of the Royal Society of Biological Sciences, March 2008, 278, 623-628. Our thanks to the researchers, Dr. David Coltman (Professor) and his PhD student, Jocelyn Poissant for their help on this article. Dr Coltman also recognizes the work of his collaborator Professor Marco Festa-Bianchet.

Often the natural benefits are ignored when land is used for “a more economic” purposes, such as infrastructure, extraction, urban development or even agriculture, largely because the ecosystem services are either not valued or difficult to value at a market level. Typical examples of this would include the unsustainable use of land, such as deforestation where land owners don’t always realise the long term dangers of lost forest services. Everything from unregulated urban or coastal development, over fishing, unsustainable agriculture and extractive industries are affecting the biodiversity of local ecosystems. The issue being that many of the ecosystem services directly benefit human welfare; from water supply, cleaner air to food, so if these systems are unbalanced they are likely to negatively impact on local human welfare and health.

The PES system has been developed where the beneficiaries of the ecosystems pay the providers of the services (i.e. the landowners) to protect their local ecosystem. Examples of this might be tourist companies paying African communities for the protection of local wildlife or a water company buying watershed protection from farmers near the water source. It is vitally important that land owners recognise the importance of their local ecosystems and its services, and help to protect it, even if only for their own well-being and personal future profitability.

Normal definitions of PES are based around the idea of one ecosystem provider and one buyer of the services, and often auctions or competitive tenders are used to efficiently allocate the resources for ecosystem services. However, it’s not news that the best result for improved or protected ecosystems is with a larger area, but the key to this is encouraging the landowners to collaborate with their ecosystem services tenders, so that these services can be provided at a landscape scale. But it isn’t just about the total number of ecosystem sites, it’s also the location within the landscape relative to other sites, because biodiversity conservation tends to be more effective when conserved sites are connected together to form a corridor or larger habitat area.

In order to find a way to encourage landowners to submit bids that will work together for a more productive and larger ecosystem, Dr Andrew Reeson and a group of researchers at CSIRO Sustainable Ecosystems, Australia have studied how best to create a collaborative auction system on a landscape scale.

A landscape scale based ecosystem service needs to cover a region of interconnected areas, the efforts need to address specific conservation objectives and the landowners in that region need to collaborate to achieve those objectives. The actual value of each bid within this system is therefore dependent on the other bids in the final successful group. There are then 2 key issues to cover in such instances, one being the combined value of the ecosystem services and second to coordinate the individual landowners to offer the best group of bids. Iterative auctions seem to offer the best hope as they allow for information and strategies on each bidder to be made available as the auctions continue, as it encourages landowners to submit bids which align with their neighbours. But it doesn’t always work, as coordination may in fact lead to collusion by landowners.

The group of Australian researchers decided to set up auction trials to find out if collaboration could be encouraged. A series of experiments was run in which participants took on the role of landowners in a conservation auction. To make the scenario more realistic they received payments based on the results of the auction and their land management decisions. This approach allows a number of variations of the auction mechanism to be tested and compared under laboratory conditions. GenStat’s generalized linear models (GLM) were used to analyse individual bidding behaviour and simulate biodiversity outcomes in the auctions.

One measure of the effectiveness of an auction is the amount by which individuals inflate their bids – the higher their bid prices, the less land can be secured for conservation. GenStat’s GLM tested how bid prices were impacted by various auction rules. The results showed that in auctions with more rounds participants initially inflated their prices by more; however where the number of rounds was unknown to participants, this strategic behaviour was greatly reduced, resulting in a more efficient auction. It appears that participants didn’t dare inflate their bids too far in case of missing out if the auction suddenly ended.

Another test is the amount of ecosystem services provided across the simulated landscape in the auction. Analysis of a GLM showed that this was higher in iterated auctions than for one-off auctions, as there were more opportunities for landowners to link up with their neighbours to form conservation corridors. The analysis also showed that iterated auctions were more effective when some bids were locked in early on. These results show that iterated auctions have the potential to deliver landscape-scale ecosystem services, such as connected conservation areas, in an efficient manner. This can be applied by conservation managers to ensure they get the best possible bang for the limited bucks they have available to reconnect fragmented landscapes.

Find out more about the project, or to read the full paper, go here.

Our thanks to Dr Andrew Reeson and colleagues for their help and collaboration with this.

A key user of GenStat is the IAMZ in Zaragoza, Spain, whose main aim is to train people in how to do research. The IAMZ stands (In Spanish) for the Mediterranean Agronomic Institute of Zaragoza, it is one of the 4 sites of the CIHEAM (International Centre for Advanced Mediterranean Agronomic Studies). The main purpose of the CIHEAM is “to develop cooperation between Mediterranean countries through postgraduate training and promotion of cooperative research in the field of agriculture and natural resources.” As a part of this the IMAZ develops Master programmes and advanced courses for professionals designed to update and further the training of researchers in agriculture and its related sciences.

Students come from all over the world to study for the Master programmes, some are straight from their first degree and wish to complete their education, and some are already at work and wish to update their knowledge and understanding of agricultural research. I spoke with Dr Ramzi Belkhodja from the IMAZ, who heads up the Master in plant breeding. The programme covers several areas, one of which being statistics and the use of statistics in plant breeding research. Dr Belkhodja and his colleagues recognised the need for anyone either in plant breeding research or planning on entering the field, to have a good understanding and grounding in statistical methods.

Currently the team teach their statistics course using GenStat. Given GenStat’s history in agricultural research it is perhaps no surprise that it is used as the phraseology and terminology reflects the language used in the biosciences. GenStat contains a huge range of statistical techniques; in fact you would be hard pushed to find a statistical test or routine not included in GenStat. However, it isn’t just GenStat’s inherent statistical strength or its history that makes it so useful to Dr Belkhodja and his colleagues.

GenStat’s straightforward Windows menu system means that the statistics is easy to teach and importantly easy for the students to pick up and understand, indeed Dr Belkhodja told me that the students are up and running with GenStat and statistical techniques within 3 teaching sessions. The self-explanatory menu system means that the statistical methods and concepts are easy to manage and understand. Combined with the speed and ease of importing any data from Excel (for example) into GenStat, and the lecturers at IMAZ can spend less time on the complicated syntax and more time on the analysis itself.

Given the students come from all over the world; GenStat also provides an added advantage for those coming from the developing world. When the students return, they can obtain a free copy of GenStat – as GenStat Discovery, to continue their research. So, in a world where “time is money”, students don’t need to learn a new package in their work environment; they can implement their learning and statistical understanding in a package they already know and trust.

Researchers at the New South Wales Department of Primary Industry in Australia highlight the complexities of farming using one of the more traditional methods of crop production. Southern Australian farms have long used annual legumes in the pasture phase of crop rotation methods. These act as natural disease breaks and restore soil structure and nitrogen in the soil, as well as providing valuable, high quality feed for livestock. Fantastic! In the panic of today’s world on pesticide poisoning this seems an ideal solution, and has certainly been a good one – here we are using the earth’s natural resources and traditional crop rotation methods to help increase production, and as a bonus create livestock feed…however it is not that straightforward.

The problem with annual pastures and crops, is that they are often shallow rooted and therefore use less rainfall than the deeper rooted perennial species. This has led to a significant increase in groundwater recharge with higher water tables and dry land salinisation in parts of Australia where the agricultural production has meant a change from perennial deep-rooted species to annual crops and pastures. Simple answer! Surely we just swap the annual pastures and crops for perennial ones… but which one? Australia is a big country with a diverse range of climates. How then do you determine which perennial legumes and herbs are best suited to Australian conditions, with so much variation between regions?

A team of researchers headed by Guangdi Li took on the challenge. The purpose of their study was to evaluate a wide range of herbaceous perennial species as potential new species for incorporation into the farming system of Southern Australia. Guangdi Li’s team needed to identify species with an ability to adapt to a broad range of environments, and species which are particularly suited to specific soil conditions, environments and climatic patterns. They evaluated 91 perennial legumes and herbs (entries) at 10 sites across Southern Australia. There were originally 17 sites but 7 were abandoned because of weeds or poor establishment leaving 4 in Western Australia, 1 in South Australia and 5 sites in New South Wales. Only a subset of the 91 was sown at each site, local conditions not being conducive to the successful establishment of many of the varieties. In fact only a handful of the entries were sown at all the sites. Some of the sites were chosen because they posed environmental constraints to plant growth and hence could be used to identify alternate species that may be better adapted to these more demanding environments. Those constraints were saline or waterlogged soils, although drier than usual conditions minimised the effects of waterlogging. Over the three years of the experiment there were a number of occasions that herbage mass and plant frequency was assessed across the ten sites, creating 67 and 21 ‘environments’. Each environment provides information about the performance of the varieties, but of course successive samplings from the same site are not independent of each other, and a high correlation was expected.

ASReml was chosen to analyse this data because it can handle the estimation of the genetic correlations between pairs of environments. The analysis involves a mixed model in which the variance-covariance matrix of the entries in different environments is modelled using a factor analytic (FA) structure. The FA model facilitates the accurate prediction of entry means for individual environments using best linear unbiased prediction. Selection across a range of environments, be it the saline or waterlogged sites or even for those samplings taken in Summer for instance, is easily achieved by using the Predict facility in ASReml. Entries that showed potential in saline or waterlogged conditions can be targeted for inclusion in future breeding programs.

The complexities of this study very much reflect the complexities faced by the agricultural producers.  In farming there are often no straightforward, obvious answers to the questions posed and the problems faced. This is why research and studies such as Guandi Li’s are so important; they give evidence based on sound science and statistical analysis techniques which show the importance of specific types of plants; in this instance Lucerne, which performed well over a broad range of environments. This in itself suggests projects are needed to fully exploit this plant so that its limiting factors (susceptibility to acid conditions and heavy grazing) can be overcome. Additionally Guangdi Li and his team showed that there is a range of deep-rooted perennial legumes and herbs which could be used in the Southern Australian farming system, rather than relying on a narrow range. The impact of the results of this and future trials are huge, it is yet another step closer to even better planned farming policies and methods, and hence to a true understanding of how we best feed ourselves and the world.

There is no doubt that sound scientific research is needed on which to base future crop, farming methods or policy decisions; research that is built using proven scientific tools. VSNi are a key provider of these tools; we know this because many agricultural scientists rely on us to continue providing high quality, relevant statistical analysis software built by statisticians who understand the complexities of agricultural research and data.

Anyone who loves Italian food is likely to have a strong opinion about the best texture of cooked pasta. But how can you gauge this?

And therefore give real, practical cooking tips? Or even know what type of pasta you prefer? Particularly as there are so many different factors affecting the texture of cooked pasta: length of cooking time, length of time left standing, size and type of pasta, as well as the brand itself. I know when I was a student the best method was always considered to be throwing a strand of spaghetti at the wall and seeing if it stuck; but looking back I think this method was more for the food fights which usually followed. Thankfully the food manufacturers have developed a more scientific method for testing the firmness of pasta, but even this is fraught with complications.

Until recently using trained panellists was considered the best tool for measuring the cooking quality of pasta products; however this is open to bias; be it localised preferences, leading to the prevention of an internationally recognised standard for testing pasta texture, or just personal, individual preferences. Using panellists is also expensive, slow and needs a large sample. Over time various different methods and instruments have been developed to evaluate pasta. However, it seems that different cooking and instrument settings in different laboratories have influenced sample rankings within the tests. These issues became apparent to Dr Mike Sissons and his colleagues, who are part of the Durum Wheat breeding program at Tamworth Agricultural Institute, in Australia. One of Mike’s jobs is to test the firmness of the Durum wheat which has come from varieties trials. To do this he has to make the spaghetti, dry it and then cook it, but the firmness is very sensitive to many of the choices made in how to test the spaghetti firmness. How many strands to test, how are they arranged, what is the speed of the cut etc. The biggest factor is over and under cooking and the continued cooking of the spaghetti after it has been removed from the water.

Mike’s experiments were to determine the effect on firmness of changing each of these factors. Differences of up to 10% were noticed between the firmness readings at Tamworth and in a collaborating laboratory in Canada. These differences far exceed differences between varieties and a standard method of testing of spaghetti firmness was desired.

Happily a paper recently published at www.redorbit.com covers this very point. The study was designed to investigate the effect of process and instrument variables on the testing of the firmness of cooked pasta and from this to devise a standard procedure that can be replicated in laboratories so that true differences can be seen between similar examples of pasta.

Three commercial spaghetti samples were used, with different strand diameters and cooked in the same way; then the texture was analysed in such a way to provide 12 different measurements per sample. Different factors such as spaghetti weight, length of strand, cooling water temperature etc were varied; this allowed for researchers to find out the best conditions for testing the pasta’s firmness. The researchers then used the linear mixed models procedures available in ASReml to analyse the firmness of the spaghetti samples.

Responses of spaghetti firmness to most of the factors was linear or could be made linear with a transformation so the fitting of a model was very simple. However the advantage with ASReml is that you can add complex variance structures to a model which often cannot be done with other packages. In this simple case ASReml allows the researchers to estimate the variance for each level of factor in each experiment and determine if any of the levels had significantly lower variance. If the model suggested differing variances the level of factor with the lower variance was recommended as the setting to use. With both the Tamworth and Canadian laboratories using a standard method the correlations between laboratories was greatly improved and average differences between the laboratories was reduced.

This study has shown that there are ways of increasing the repeatability of the results, and hence make it easier to discriminate between textural differences in different pasta samples, and shows there is an alternative to sensory analysis. The impact for laboratories, the Durum trade and pasta manufacturing industry is huge; decisions on wheat variety, ingredients, recommended cooking times etc are based on science not taste buds.

As a footnote - throwing spaghetti at the wall is still used as a valid test of stickiness, for me personally this means I can happily continue trying to redecorate my kitchen!

In Australia the Grains Research and Development Corporation, in conjunction with the Australian Crop Accreditation System have set up the National Variety Trials (NVT) program. A program designed to provide information on newly released crop varieties to aid with crop variety selection decisions, based upon their individual growing conditions. Driven by the increasing commercial focus of plant breeding programs in Australia the NVT program conducts around 600 trials on the full range of commercially significant field crops including wheat, barley, triticale, oat, canola, lupin, lentil, field pea, faba bean and chickpea in over 250 geographic distinct locations.  This national program of comparative crop variety testing provides standardised trial management, data generation, collection and dissemination and is managed through an internet accessed database, that ensures a common approach and uniformity across the system.

The NVT program allows for single national analysis for each crop rather than state- based ones, and the results of the analysis are presented in terms of an estimated yielding ability for a specific environment. So whilst the data collected is nationally, providing more data and information and allowing for greater predictability, the results are now provided for individual areas and environmental conditions.

The large number of trials and consequently huge amounts of data (for example 210,000 records for wheat alone)  collected needs a data analysis system that can manage large datasets; and given that the results are used in choosing grain varieties, which will have an economic impact, such a tool must be highly accurate and trusted. The biometricians who run the analysis use ASReml.

The analysis model which is used is unique to ASReml as it involves the use of a two-stage linear mixed model in which the data from each trial is weighted according to its statistical and biological reliability. The linear mixed model includes terms which aim to model the variety by environment interactions in a plausible and interpretable manner.

”ASReml is the package that helps farmers, breeders and crop variety evaluators obtain the most reliable predictions of genetic value for a range of crops grown in different environments; farmers can get the best information available about performance of varieties in their own location and make an informed decision,” says Professor Brian Cullis, Research Leader for DPI Biometrics and leader of the SAGI (Statistics for the Australian Grains Industry) project, the project which run the analyses for the NVT program.  “ASReml is the only package we trust to run this type of analysis because of its proven accuracy, speed and flexibility for these types of complex two-stage models.”

ASReml is a vital tool in the NVT program; it has allowed for the analysis of huge amounts of data on the performance of different crops under different conditions. Without reliable analysis tools the reports would not have the credibility they have, and therefore Australian growers would not have the access to such useful reports to aid them in crop selection.