By Emily Lehmann
Oscar hype is in full-swing, and we all have our favourites for Hollywood’s night of nights (we must admit we are partial to Birdman taking home the ‘best picture’ gong). But the big-screen isn’t the only place to find world-class movies.
At our Discovery Centre in Canberra yesterday, we unveiled two world-class movies of our own. The animations, created by up-and-coming Australian biomedical animators, uses the latest data visualisation techniques to bring science to life in incredible 3D detail.
Created by Australian up-and-coming biomedical animators using the latest data visualisation techniques, they feature key research into Alzheimer’s disease and type 2 diabetes from CSIRO and the Walter and Eliza Hall Institute of Medical Research (WEHI).
Through narrated picture, the animations explain very complex biological processes related to each disease with scientific accuracy: zooming in on what happens inside our body but can’t be seen with the naked eye.
The animations illustrate key research techniques into Alzheimer’s disease and type 2 diabetes, based on work we have done with the Walter and Eliza Hall Institute of Medical Research (WEHI).
The first video looks at Alzheimer’s disease – the most common form of dementia – which affects one in four people over the age of 85, a number that will increase significantly as our population ages.
This animation takes you on a journey to the neurons of the human brain, revealing how normal protein breakdown processes become dysfunctional, and cause plaque to form during Alzheimer’s disease.
This build up of plaque in the brain can take decades and is one of the main indicators of the disease.
The Insulin Receptor and Type 2 Diabetes
About one million Australians currently live with diabetes and about 100,000 new diagnoses are made each year.
These staggering statistics are fuelling research efforts aimed at finding a cure or ways to prevent or better manage the disease.
Highlighting a recent discovery by WEHI, this animation focuses on the role that the insulin receptors play in the disease and what might cause resistance to the hormone insulin.
It’s part two in a series of animations on type 2 diabetes, you can check out part one here.
These are the second round of animations created through VizbiPlus – a joint project between CSIRO, WEHI and the Garvan Institute of Medical Research.
Under the guidance of internationally-acclaimed biomedical animator Drew Berry from WEHI, VizbiPlus is training-up the next generation of biomedical animators, to raise the bar in science communication and bring critical research to the world.
You can read more about our data visualisation work here.
Heart rhythm disease is a life-threatening, electrical disorder that stops the heart from pumping blood effectively. It is a lethal condition that is responsible for around 12 per cent of Australian deaths each year.
In order to open the door to better diagnosis and treatment for heart rhythm disease, we’ve been working with the Victor Chang Cardiac Research Institute to develop our very own ‘virtual heart’. What’s more, we’ve done this using the same technology that drives your favourite computer games.
Impressively, when we ran a simulation through the virtual heart, it was able to model hundreds of thousands of different heart beats. This then allowed scientists to screen all of those heart beats, and search for abnormalities.
According to the Victor Chang Institute’s Dr Adam Hill who led the research, this has taken us a step closer to understanding rhythm disturbances in our most vital muscle.
“This research is hugely exciting! We were able to identify why some patients have abnormal ECG signals, and how a person’s genetic background can affect the severity of their disease,” he says.
Analysis on this scale has simply never been possible before. The simulation took just ten days, thanks to the computational grunt of CSIRO’s Bragg supercomputer cluster, which combines traditional CPUs with more powerful graphics processing units or GPUs.
GPUs have typically been used to render complex graphics in computer games. However they can also be used to accelerate scientific computing by multi-tasking on hundreds of computing cores.
By comparison, if you were to try to do the same simulation using a standard desktop PC, it would take 21 years to get the job done.
Adam hopes the new technology will help doctors read ECGs more accurately, which will mean faster, more accurate diagnosis of heart rhythm disease. By understanding why the same disorder affects people differently, the right treatment can be given to the right patients.
Scientists at the Victor Chang Institute are now using these discoveries to develop automatic computerised tools for diagnosing heart rhythm disorders.
Read more about how we’re using data and digital technologies to tackle health challenges on our website.
By Andrew Warren
If you’re a regular at the gym or an early morning boot-camp fanatic, it’s possible that the first thing you picture when you think of protein is the powder you use to make your post-workout recovery shake.
But when our scientists discuss protein, they’re talking about the many thousands of molecules that act as the essential building blocks of life as we know it. Because proteins are so important to constructing life, researchers need a way to visualise the exact ways in which they fit together so that they can better understand the functions they play in our bodies.
With this in mind, a team of international programmers and bioinformaticians (think biology, computer science and maths mixed together) led by our very own Dr Seán O’Donoghue have created a new web-based tool named Aquaria that can create unprecedented 3D representations of protein structures.
Aquaria is based on the Protein Data Bank, an online resource which houses more than 100,000 structures of proteins that contains a wealth of detail about the molecular processes of life. But Sean and his teams were conscious that few biologists were taking full advantage of the site. The Protein Data Bank is designed for and by biologists who are expert in structures; however for most biologists, its organisation can be confusing.
So, they created Aquaria to make this valuable information more accessible and easier to use for discovery purposes.
Freely and publicly accessible, Aquaria can help scientists like ecologists, nutritionists and agriculture, biosecurity and medical researchers to streamline their discovery process and gain new insight into protein structures.
Sean’s team added additional layers of information (like genetic differences) to the basic protein structure and made it accessible in a fast, easy-to-use interface that’s visualised in a fully 3D environment.
“We’ve added protein sequences that don’t yet have a structure, but are similar to something in the Protein Data Bank,” says Sean.
“That meant we first had to find all these similarities. We took over 500,000 protein sequences and compared every one of them with the 100,000 known protein structures, and that has given us around 46 million computer models.
“For example, you can add Single Nucleotide Polymorphisms (SNPs) that cause protein changes, then visualise exactly where those changes occur in the protein structure. This provides valuable insight into why proteins sometimes completely change their function as a result of one small change in the DNA code.
“You can then ask interesting questions like ‘Does this set of SNPs cluster in 3D?’ and the answers to such questions can set new research directions.”
Aquaria was developed in collaboration with Dr Andrea Schafferhans from the Technical University of Munich, and is hosted with support of a grant from Amazon Web Services.
To learn more about Aquaria, you can take part in a special webinar scheduled for 9am Tuesday, 3 February (AEDST).
The International Year of Crystallography is drawing to a close, and we’re not going to let it finish without showing you something about what crystallographers do. Which is not what most people would assume when they hear the word: there are crystals involved, but it’s not exactly the study of crystals as we generally think of them. It’s the study of how matter is organised, using crystals as a tool.
Now, naturally we want to know how matter is arranged. Apart from being very, very interesting to find out about, it also helps in many different fields, from drug delivery to materials science. In fact, it was crystallography that provided – controversially – the key to understanding the structure of DNA.
So assume you want to look at something in the greatest possible detail, seeing its smallest possible components. Obviously, you’d use a microscope. But there’s a limit to the smallness of things you can see that way: the wavelength of the light human eyes see. Visible light has a frequency of between roughly 400 and 700 nanometres, and can’t detect atoms, which are separated by 0.1 nanometres. This is the perfect frequency for X-rays.
We can’t make appropriate X-ray lenses to make x-ray microscopes to study molecules: we have to do it in a roundabout way. We beam X-rays onto crystals, scattering the rays, in just the same way that light reflects when it hits an object. Then we use a computer to reassemble the rays —the diffraction pattern —into an image. The diffraction of a single molecule would be so weak that we couldn’t get any meaningful information from it, so we use crystals, which have many molecules in an ordered array, to amplify the signal so we can see it. Crystals are highly ordered structures, made up of 1012 or more molecules, makes the x-ray diffraction patterns — the main tool of crystallography —possible to analyse.
Crystallographers were among the first scientists to use computers, and used them to do the advanced calculations needed to reassemble diffraction patterns into coherent images. That’s why it seemed fitting to name our supercomputer after the founders of crystallography – Lawrence and Henry Bragg. Lawrence was the first person to solve a molecular structure using x-ray diffraction.
Today we can not only view molecules in 3D, but also study the way they operate. Improvements in x-ray machines have also led to synchrotron facilities, which can produce far more efficient and precise beams.
And speaking of synchrotrons …
One of our crystallographers, Tom Peat, has deposited more than 120 structures in the Protein Data Bank using data collected at the Australian Synchrotron. They were all derived from crystals developed in CSIRO’s Collaborative Crystallisation Centre.
This is one of our favourite structures.
It’s the structure of AtzF. This enzyme forms part of the breakdown pathway for atrazine, a commonly used herbicide. We’re trying to understand enzymes better and use them for bioremediation – cleaning up environmental detritus such as pesticides and herbicides – and we’ve now solved the structures of four of the six enzymes involved in the atrazine breakdown pathway. We also look at protein engineering, to see if we can make these enzymes even more effective at cleaning up the environment.
Before we get to the crystal image, there are other steps on the way. First, someone has to grow the crystals (clone the protein, express it, purify it and crystallise it). Then it’s off to the Synchrotron to get a data set (many diffraction images in sequence). Here you can see an actual protein crystal.
The picture on the right is the diffraction image.
The crystallographers measure the intensity of the reflections (the dark dots). They combine that with the geometry and use some complicated maths (a Fourier Transform) to produce an electron density map. They then use that map to build a model.
Not all our crystallography work is in the same area. We also work on some pharmaceutical applications. One of our projects, with hugely important implications for human health, is on the design of desperately needed new antibiotics. We’ve been collaborating with Monash University, looking at the pathway that sulpha drugs (such as sulfamethoxazole)– the ones we used to treat bacterial infections prior to the discovery of penicillin – take to treat Golden Staph infections in humans. The aim is to design new antibiotics that target the same pathway. You can read a paper that describes our recent findings in the Journal of Medicinal Chemistry, and here’s a picture of what we’ve been doing.
We think this deserves its own Year. And we hope it’s clear just how important it is. Crystal clear.
Australia’s “ferals” — invasive alien plants, pests and diseases — are the largest bioeconomic threats to Australian agriculture. They also harm our natural ecosystems and biodiversity. Some, such as mosquitoes, also act as carriers of human diseases.
One method of controlling invasive plants and pests — known as biological control, or “biocontrol”— is to use their own enemies against them. These “biocontrol agents” can be bacteria, fungi, viruses, or parasitic or predatory organisms, such as insects.
To find biocontrol agents, we travel to the native home of invasive species and search for suitable natural enemies. After extensive safety testing, they are introduced into Australia.
But do they work?
Learning from the cane toad catastrophe
Cane toads, which were introduced in 1935 to control cane beetles in Queensland’s sugar cane crops, are probably the most infamous example of biocontrol going wrong in Australia.
But Australia’s borders were more open back then. To protect against such harmful mistakes, Australia now has world-leading biosecurity import regulations and an effective quarantine system.
To be allowed entry into Australia, a candidate biocontrol agent must be assessed using internationally-recognised protocols. This demonstrates that it will not pose unacceptable risks to domestic, agricultural, and native species.
A cost-effective solution
Other control methods, such as the use of poisons and mechanical removal, require continued reapplication. Many biocontrol agents of plants and insects, once established, are self-sustaining and don’t have to be reapplied.
Prickly pear is a perfect success story of biocontrol. The plant was introduced into Australia in the late 1770s and grown in a few areas of NSW and Queensland until it became invasive after rapidly spreading following the flood of 1893. Biocontrol was initiated in the early 1900s and the prickly pear moth, Cactoblastis cactorum, was introduced in 1926 from the pear’s native home in the Americas. Cactoblastis has been keeping prickly pear under control almost by itself to this day.
Since then, many more biocontrol agents have been introduced to control invasive plants. These include mimosa in our top end, bridal creeper in southern Australia, parthenium in Queensland and ragwort in Tasmania.
A series of cost-benefit analyses in 2006 revealed that for every dollar spent on biocontrol of invasive plants, agricultural industries and society benefited by A$23. This was due to increases in production, multi-billion dollar savings in control costs and benefits to human health.
Biocontrol has also proven to be the only effective way to significantly reduce European rabbits across Australia. Myxoma virus was released in 1950, followed by rabbit calicivirus in 1995, causing regular disease outbreaks in wild rabbits. Together, they have kept rabbit numbers well below the devastating pre-1950s levels.
It’s estimated that the benefit of rabbit biocontrol to agriculture is worth more than A$70 billion. This is the only example of a successful large-scale biocontrol program against a vertebrate pest anywhere in the world.
The initial costs of biocontrol programs are generally high. That’s because we have to find suitable candidate agents overseas, test them for safety in quarantine, and comply with regulations around release.
But once biocontrol agents are released and affect the invasive species across its range, follow up control costs are greatly reduced.
Biocontrol is not a ‘silver bullet’
Biocontrol will not solve all problems to do with invasive species.
Weather and climate can affect biocontrol agents, like all living organisms. These two factors can slow and even stop the agents building-up to sufficient levels to control the invasive species.
In the case of the two rabbit viruses, virus-host co-evolution has led to a decline in effectiveness of the viruses over time as they lost virulence and rabbits developed resistance to them. This is similar to how bacteria can develop resistance to antibiotics. As a result, we must continue to search for ways to counteract these effects.
Like a multi-drug cocktail, biocontrol agents must often be used together to knock out an invasive species. And while biocontrol rarely completely eradicates an invasive species on its own, it may control it enough to be able to use other methods at a lower cost.
And just because we use biocontrol, it doesn’t mean we don’t need good farm practices and land management, such as bush restoration, to ensure the recovery of ecosystems affected by invasive species.
Biocontrol is unlikely to be the solution where invasive species are very closely related to species that we value — cats, for instance. Feral cats have recently been in the media as the greatest threat to Australia’s mammals. But because they are the same species as the cherished family moggy, a biocontrol program would be highly controversial.
New biocontrol programs
The historic successes of biocontrol in Australia justify continued investment. For widespread invasive species, there are no alternatives as cost-effective that work across the vast landscapes where feral species roam.
For example, the European carp pest makes up 90% of the fish biomass in the Murray Darling river system. The most promising option being developed for large-scale control is the carp-specific koi-herpes virus that is in the final stages of testing (to make sure the virus only targets carp). Its proposed release in Australia will soon be open for public debate.
Another case is the recent release of a rust fungus from Mexico for the biocontrol of crofton weed in eastern Australia. This invasive plant smothers grazing systems and natural ecosystems, including on the hillsides of Lord Howe Island, a World Heritage Area. The expectation is that this new highly-specific rust fungus will significantly contribute to control of this plant, the way other rust fungi have successfully done in the past against other invasive plants.
After 100 years of history in Australia, biocontrol should continue to have a bright future given it is the only approach that is environmentally-friendly, cheap and effective.
This article was originally published on The Conversation.
It’s a hard life being a small farmer in sub-Saharan Africa. About 200 million people in the region are poor and undernourished. Most of them are smallholder farmers in rural areas, who rely on agriculture as their main source of food and income.
Part of the reason for their level of hardship is that the major staple crops, sorghum and cowpeas (which provide not just food but fuel and fodder for livestock) have low yields. Poor soil, low-quality seed, drought and disease all play their part.
Obviously, if these farmers could get greater productivity from their crops, they could have a secure supply of food, and possibly even be able to sell the excess and bring in some extra money. But conventional genetic improvement to increase yield is a slow process, and these farmers are hungry now.
So: how to make improved yields happen?
The Bill and Melinda Gates Foundation has just awarded us a $14.5 million grant to work on it. The five year project, in partnership with other world leading research teams from Switzerland, USA, Germany and Mexico, will develop tools to generate self-reproducing hybrid cowpea and sorghum crops.
What we’re planning to do is to develop high-yielding sorghum and cowpea crops that have seeds the farmers can save and grow, and which don’t decrease in quality or yield. And that’s going to mean making a very fundamental change to the way they’re bred – changing from sexual reproduction to asexual.
Hybrid crops can produce yield increases of 30 per cent or more, because of what’s known as hybrid vigour – basically that some crosses between two strains of crop will combine the favourable traits of both parents and be more successful than either.
Hybrid vigour is the same mechanism which produces the loveable labradoodle. A labradoodle puppy inherits the favourable traits from its purebred Labrador and poodle parents. However, two labradoodles won’t produce labradoodle puppies (they’ll be more Labrador-ish, or more poodle-ish). In just the same way, the seed from hybrid crops will not express the favourable traits. The puppy’s increased adorableness is of course a matter of personal opinion but it is a furry demonstration of hybrid vigour.
Unfortunately, current technologies to produce hybrid seed (and labradoodles) are expensive, and farmers need to buy new seed every year as the favourable traits only last one generation.
If we can develop self-reproducing hybrid cowpea and sorghum crops the farmers would then be able to self-harvest high-quality seed, giving them a more secure food supply and possibly even increased income from selling excess seed.
It’s a big challenge. As project leader Dr Anna Koltunow explains ‘It’s not going to be easy, otherwise it would have been done already. The idea of changing the plants’ reproductive process to an asexual one is a complex undertaking’.
The first stage of the project will involve developing the techniques that will allow cowpea and sorghum plants to reproduce asexually. This is lab-based work. If this stage is successful, African breeders and institutes will join the project for the subsequent phases.
Biological illustration has come on a bit from the days of Gould’s gorgeous illustrations of birds, or Leonardo’s Vitruvian Man. Today, with the help of big data and big graphics power, we can visualise things, not just at the molecular level, but at work.
But why – apart from because it’s beautiful and fascinating – do we do it? How is it helpful? What can it show us?
Obviously, we’ve been using rudimentary data visualisation for a very long time. Charts, maps, tables, graphs. All data visualisations, but not at the level we now find ourselves working at. As Sean O’Donoghue, from our Digital Productivity and Services Flagship, puts it, ‘Data visualisation is a new visual language; we need to become fluent in it to manage the complexity of computational biology’.
Let’s think about genomic data. The more we know, the more we need new tools to deal with the knowledge we have. And we now know a lot. We’ve got the ability to generate tremendous amounts of genomic data from sequencing. Analysing that data is now the roadblock to our being able to convert what we’ve found into something useable.
Obviously, some genome analysis can be done using automated processes. But that still leaves a lot that depends on human judgement, particularly in the early stages such as hypothesis formation. Our concentration – and eyes – frankly aren’t up to spotting something different in a field of As, Cs, Gs and Ts (and nothing else), that seems to go on forever. Think of Where’s Wally?, in monochrome, with one Wally hidden on a single page hundreds of times larger than book pages. And then imagine that finding the Wally you’re looking for could make a big difference to people’s lives.
If we can combine visual and automated analysis, the pairing becomes more powerful. Users can seamlessly look at their data and perform computations on it, refining their analysis with each step.
Visualising also helps us reason about complex data. Sometimes, a well-chosen visualisation can make the solution to a complex problem immediately obvious. That’s because of the way that visual representations simultaneously engage the eyes and the memory. When we look at a visual, our eyes and our brain work in parallel to take in new information, and break it into small chunks. Then both the eyes and the brain process the bits in their different ways to extract meaning. It works like this.
You’ve gone to the supermarket – not your usual one – to buy bananas. When you walk in, your eyes scan the layout. At the same time, your brain is processing the various sections of the layout, and telling your eyes to home in on the fruit and veg section. It does this by sending signals from memory about how fruits look. Your eyes then break the entire scanned area into parts, then scan each part until they (all but instantly) recognise the veggie section. The same process is repeated until you spot the bananas in the fruits section. Your eyes and memory do their own things but work in parallel.
We’ve used our brains to build tools that can help us discover more and more. But making sense of what they’ve found still depends on us and our limitations. Around half of the human brain is devoted – directly or indirectly – to vision. Visualising the vast streams of data lets us work with what we’ve got to make it something more than a hunt for a tiny needle in a monstrous haystack.
If you want to see more data visualisations, there are some beautiful ones at Vizbi.