Never before has splitting hairs been so useful. Analysing the proteins in hair could provide a new way to identify people when DNA sequencing fails.

DNA profiling is often used to identify criminals or archaeological remains, but has its problems. While DNA can last for hundreds of thousands of years under some conditions, it breaks down quickly when exposed to water, light or heat. In 2009, this prompted the US National Research Council to issue a call for highly reliable forensic identification tests that can be used when DNA isn’t up to the job.

Glendon Parker at the Lawrence Livermore National Laboratory in California and his team have turned to hair, which we shed from our bodies and remains relatively well preserved over time.

Hairs are made of proteins that are pieced together using genetic instructions in our DNA. When mutations occur in these genes, it can lead to slight changes in the sequence of amino acid molecules that make up these proteins.

To use such changes to identify different people, the team have developed a technique for breaking down a hair’s proteins and separating out the resulting chunks to analyse their sequences. Comparing these results to a database of hair proteins can identify places where unexpected amino acids have been inserted.

Crucial leads

Using this approach, the team were able to identify dozens of sequence changes in the hair of 76 living people, and hair from people who died around 1750 to 1850. They calculated that some of the changes they found could be as rare as occurring in only one in every 12,500 people, making them powerful markers for linking a criminal’s head to the hairs left at the scene of a crime.

“Because protein is more abundant and more robust than DNA, this potentially opens up enormous avenues of research in bioarchaeology and forensic science that couldn’t have been tackled before,” says Andrew Wilson at the University of Bradford, UK, who provided the old hair samples used to test the technique. It could help answer difficult questions, like whether people found at archaeological sites are related, and where they may have travelled from, he says.

“It has a remarkable potential to provide investigators with crucial leads,” says , at Sheffield Hallam University, UK.

One drawback is that the technique requires fairly large samples of hair, and the team are now working to reduce the amount needed to do this kind of analysis. That’s important because a single hair is usually all that is found at a crime scene, says Francese.

Materials called perovskites will transform the way we build and use solar cells – by making them bendy, says the researcher behind the idea

What’s the special ingredient in the solar cells you’re designing?
Perovskites are a large family of materials with a common crystalline structure, and include natural minerals or films synthesised in the lab. Today most solar cells are made of silicon, which has to be at least 80 micrometres thick – about the width of a human hair – to convert as much light energy as possible into electricity. Perovskite cells can be just tenths of a micrometre thick, making them ultra-lightweight and very cheap, while still keeping the efficiency of bulkier silicon cells.

What can they do that silicon cells can’t?
Our vision for them hinges on the fact that they can be flexible because they’re so thin. They can be printed or sprayed on foils, and because the cells are only 0.2 micrometres thick, they’d be light as air. And they can be semi-transparent or coloured, so you could put them in windows, or have a building that from a distance you don’t see is covered with solar cells but that could produce enough electricity to power itself.

The best silicon cells convert about a quarter of the solar energy falling on them into electricity. Are perovskite cells as good?
The theoretical limit for perovskite cells is about 30 per cent, so at least 28 per cent should be possible. They are already at 18.4 per cent. You don’t need complicated equipment to produce these and the materials aren’t expensive. So now, everybody’s started to work on perovskites and progress is much faster.

So could perovskite replace silicon?
Silicon lasts 25 or 30 years; so far perovskites have only reached 1000 hours. Fighting silicon would be suicide, and you’d lose before you even start because of the lifetime.

Trees get energy from leaves with only about 3 per cent efficiency, but they have a lot of leaves and they are very “cheap” to produce. Every year they are exchanged for new ones. My dream is to have solar cells that after some time will degrade like leaves on a tree. Perovskites are partially organic, so it should be possible. If we can get rid of the small amount of lead they contain, it would be easier to make them safely disposable. Then you could use the cell for for power when you’re camping, for instance, and throw it away after.

Are you at all worried that the potential of perovskites might be overhyped?
Perovskites have already proved that they work, so I don’t think this will implode. I really think we can improve them, and even if we can’t they are already good enough to use where you can’t use silicon. When I look around, I imagine covering everything with perovskites.

Could your work help unearth other promising materials?
The particular perovskite we’re using now is one IBM was testing for transistors and other applications 20 years ago. The only thing they didn’t try was solar cells. If they had, we might be living in a completely different world. I think there might be a lot more materials like that out there.

This article appeared in print under the headline “Solar’s foldable future”

By studying metabolites in urine samples from 1559 US citizens, researchers identified differences between the two groups. As well as giving urine samples, the participants were asked questions about their diet and had their blood pressure measured four times over three weeks. The researchers then sought links between the metabolites, the foods the participants said they ate and their blood pressure.

They found that diet could explain between a third and a half of the blood pressure difference seen between men in the two groups, and a quarter in women. Most of this is down to different food choices.

Compared with whites, African Americans consumed less of a number of key nutrients that could potentially lower blood pressure, including vegetable protein, iron, potassium and calcium. African Americans also consumed more of some substances known to raise blood pressure, including cholesterol and glycine, an amino acid found at higher levels in meats.

Co-author Jeremy Nicholson of Imperial College London says a genetic mechanism may be amplifying the effects of bad diet. “Microscopic variations in the genome, none of which are significant in their own right, change the risk of populations when they have risky behaviours [such as eating unhealthy food],” he says.

Journal reference:

The drive to bring eco-friendly LED lighting into our homes is being stopped in its tracks by an embarrassing problem known as droop – the disappointing reduction in efficiency that happens when the light bulbs operate at the high power levels they need to shine at their brightest.

“Efficiency droop is one of the main obstacles to achieving cost-effective and high-efficiency LEDs,” says at the Gwangju Institute of Science and Technology (GIST) in South Korea. “Droop becomes a very important issue as LEDs expand into applications like [indoor] lighting where they operate at high currents.”

For years, LED production has grown in tandem with the cellphone, providing the backlight for their displays. But manufacturers will have to tackle droop before high-power LEDs can hit the big time.

The cause of LED droop is disputed, making the solution to the problem far from clear – but now, Park and colleagues at GIST have teamed up with to prop up this flagging performance with an unconventional device design.

A standard LED has a surplus of electrons on one side and a dearth of electrons – or an abundance of – on the other. Plug the LED into a circuit, and the electrons and holes , combine, and release energy as light.

Droop means that the proportion of the recombinations that produce light peaks at low electrical powers, with the record-holding prototype devices reaching about 250 lumens per watt. Raise the power to levels typically used for indoor lighting, though, and an increasing proportion of the electric current is lost as heat, so the efficiency drops below 100 lumens per watt.

A trap to catch an electron

The electrons and holes are caught in tiny traps called , where they are more likely to collide and recombine. In commercial white LEDs, quantum wells are made of (GaN) surrounded by barriers on either side made from indium gallium nitride (InGaN).

But conventional manufacturing techniques simply juxtapose the InGaN and GaN layers, creating an abrupt interface that physically strains the semiconductor material and generates an electrical field, which Park’s team suggest might cut the chances of electrons and holes combining and emitting light.

To test the idea, the researchers changed the nature of the interface between the well and its barrier by introducing the indium more gradually. That created a steady gradient between the barriers and the well. By smoothing the interface between these layers, they could reduce the strain and thus weaken the electric field surrounding the quantum wells.

When compared against a conventional design, the team’s LED rapidly becomes 20 per cent more efficient as the power goes up, generating more light and less heat. While similarly raised efficiencies have been , the Korean team’s approach reaches these levels at much lower currents and then sustains them.

“The paper appears to be an interesting contribution to the wide-ranging debate on droop,” comments from the Centre for Gallium Nitride at the University of Cambridge, UK, However, she warns that LED droop is a complicated issue and unlikely to have a single, simple solution. “To really solve this important problem will require more wide-ranging and systematic studies.”

Journal reference: