Science-- there's something for everyone

Tuesday, December 31, 2013

Just for fun: 2013 Small World Competition

I love the Nikon Small World Competition. Photographers send in images captured with light microscopes (that is, microscopes that rely on light, not ones that are easy to carry). I don't envy the judges, though, for having to choose between such amazing offerings. 

Nevertheless, they do have to make their selections. So here's the 1st prize winner, a colonial diatom Chaetoceros debilis photographed by Wim van Egmond.



Personally, I found myself mesmerized by the 16th place 'spider with the parasitic wasp on its abdomen', taken by Geir Drange.



You can see the rest of the winners, honorable mentions and images of distinction here.


Monday, December 30, 2013

Just for fun: Mars Evolution

The folks at NASA have another incredible flyover animation for us. This time, they bring us Mars from 4 billion years ago to today. And it only takes two minutes. Well worth your time.




Friday, December 27, 2013

Just for fun: Fly away parasites

I couldn't look away from this.
That's the abdomen of the wasp Polistes fuscatus from which a parasite (Xenos peckii) is emerging.



Here's a picture, by photographer Sean McCann.


Credit: Sean McCann. Polistes dorsalis, carrying a few friends



















never thought I'd feel sorry for a wasp.

More on this by Christopher Buddle at SciLogs.

Thursday, December 26, 2013

Just for fun: Best of science collection

Rather than taking the next week off, I'm going to give you guys some extra 'Just for Funs'. Enjoy!

Science Studio has collected the best science videos and podcasts of the year.

Here's just one offering: the jaw-dropping view from the International Space Station. I think the last 30 seconds are the best, but it's all amazing footage.




Tuesday, December 24, 2013

Just for fun: Human body collages

For your Christmas cheer, may I present the work of photographer Cecelia Webber?

Here's one of her pieces:

Hummingbird: Cecilia chooses her models from everyday people and refuses to edit them by changing their features. To capture the right angles she will stand on the top of tables, over railings on higher floors or balance her camera in precarious places

Take a closer look. Notice anything interesting? The entire composition is constructed by editing together pictures of nude models. The photos are digitally cut, rotated, resized and colored to suit Webber's needs.

Here's another example:

These amazing images of colourful birds and insects have been given a truly human touch


You can see more of Webber's work and read about her technique here.


Monday, December 23, 2013

Painted mice don't accumulate liver fat

Researchers from the University of Southern Denmark have been painting mice for science.

No, not this:

Fall Mouse
By ACEO, 2007

This:

















Why? To solve the mystery of why mice that are genetically lacking acyl CoA binding protein (ACBP) tend to accumulate fat in their livers. 

Organisms that have been bred without a specific gene (the gene might be removed or simply rendered non-functional) are referred to as ‘knock-outs’. Obviously, you can’t knock-out every gene, some are essential for life. However, where you can breed a knock-out, you have a tremendously useful model for seeing what that gene usually does. In this case, the ACBP knock-out seemed to be responsible for the changes in fat deposition.

The researchers noticed that in addition to fatty livers, the mice had greasy, tousled fur and leaky skin. That is, the animals lost excessive amounts of water through their skin. This in turn means that the mice also lose more heat than normal mice. The authors speculated that it was this extra heat loss that was causing them to store more fat than usual.

The scientists did two things to test this hypothesis. 

First, they bred some mice that were only lacking ACBP in their skin, but not in the rest of their bodies. Specifically, the animals’ livers had normal amounts of ACBP. Nonetheless, these specimens showed the same accumulation of fat in their livers as the total ACBP knock-outs.

Second, they painted the mice with a waterproof coating to prevent heat loss. At first, they used clear Vaseline, but grew concerned that the mice were absorbing fat from the Vaseline and that this might skew their results. So, the scientists switched to latex body paint. 

Regardless of body varnish, preventing heat loss through the skin reversed the liver fat accumulation in the ACBP-negative mice. 

Normally, ACBP knock-outs don’t live very long after weaning. If you choose to keep these animals as pets, make sure you refresh their body paint at regular intervals. 



Ditte Neess, Signe Bek, Maria Bloksgaard, Ann-Britt Marcher, Nils J. Færgeman, & Susanne Mandrup (2013). Delayed Hepatic Adaptation to Weaning in ACBP−/− Mice Is Caused by Disruption of the Epidermal Barrier Cell Reports DOI: 10.1016/j.celrep.2013.11.010.



Friday, December 20, 2013

In fieldwork, it pays to be nice

I once organized a scavenger hunt where various treasures were hidden throughout a public park. Unfortunately, some items were discovered and removed prematurely by local kids who weren’t part of our group. That wasn’t too big a problem for us, after all, it wasn't like years of research depended on those scavenger items remaining in place. That isn’t the case for scientific equipment left in the field.

So how do you prevent people from removing or tampering with traps, cameras or other interesting bits of equipment they stumble upon while out walking? It helps to leave the right kind of note on it.

Researchers from the Max Planck Institute for Ornithology and Ludwig Maximilians University left 60 equipment dummies half-hidden in public parks in Munich, Germany. Each item was accompanied with one of three labels (translated from German):
  • Personal:Part of my thesis – Please do not touch – Please call me if you have any questions and would like to know more...’ and a photograph of a juvenile squirrel. 
  • Neutral:Part of an experiment – Please do not touch – For information...’ and a warning sign. 
  • Threatening:Part of an experiment – Every theft will be reported to the police! – For information...’ and the note ‘GPS monitored!’.
In addition each label said ‘Property of the Max Planck Institute for Ornithology, Seewiesen’ and gave contact information.


Dummy equipment with threatening message. (Credit: © Eleftherios Bitzilekis)


The boxes were sealed with carefully placed pebbles inside them so that the researchers could tell whether they had been disturbed in any way. Each day, the scientists checked on the dummy equipment to see whether anything had been stolen, damaged, opened or moved. 

In one of the parks, the experiment had to be concluded prematurely because visitors were afraid the boxes were bombs. Such is the world we live in. But the authors got some interesting results at the other locations.

The personal note was by far the most effective in convincing people to leave the boxes alone. The threatening notes proved to be the least useful. People messed with equipment labeled in a threatening manner nearly twice as often as with any of the other boxes.

To be fair, vandalism of scientific equipment in the field is rare. In this study, the dummy boxes were interfered with only 12% of the time despite being purposefully easy to find. However, tampering does happen and researchers are keen to prevent it as much as possible. According to these data, adding a personal touch is the best way to go.

Looking at these results, I think if I were doing field work, I’d attach a note saying, ‘I’ve been a graduate student for six years. Please don’t tamper with my data so I can finally get out of here!



B.-Markus Clarin, Eleftherios Bitzilekis, Björn M. Siemers, & Holger R. Goerlitz (2013). Personal messages reduce vandalism and theft of unattended scientific equipment Methods in Ecology and Evolution : doi: 10.1111/2041-210X.12132.





Thursday, December 19, 2013

Ants weigh their options

When choosing between options, we have to weigh the importance of different factors. For example, if I’m deciding where to spend my next vacation, I have to consider cost, location, travel logistics, and many other aspects. Some of those aspects are more important  to me than others. Thus, I might be willing to compromise on my desire to go somewhere exotic if spending time with friends is a higher priority. However, if my friends become unavailable, I have to reweigh my options.

Ants also make choices, though not about whether to go on vacation. Their biggest decisions involve selecting a new nest. Temnothorax rugatulus ants live in rock crevices and prefer them to have dark interiors and small entrances.

Takao Sasaki and Stephen Pratt from Arizona State University gave ants a choice between a dark nest with a larger entrance and a nest with more internal light but a smaller entrance. In other words, the ants were asked to weigh the two attributes (light level and entry size) to decide which one mattered more to them.

Once their preferences were established, the researchers gave ants a new set of choices. Half the colonies (group A) could pick a home with the same light level and as the nest they had been ‘encouraged’ to vacate, or choose one with more light. The other colonies (group B) chose between a nest that was the same as the one they had just left or one with a larger entry hole. All the colonies picked a new nest that was identical to their old, standard, nest, rejecting the clearly inferior choices.

Finally, the ants were given their original choice back: smaller entry with more light or larger entry and less light. They were not allowed to stick with a standard nest, they had to pick one attribute over the other.

Figure 1.

Pre-treatment: An initial binary choice between sites E (small entry) and L (low light level) showed how colonies weighted entrance size and interior light level.
Light/entrance treatment: ants chose between a standard nest (S) and another that was inferior to the standard nest in one attribute, but identical to it in the other. For half the colonies (group A), the inferior attribute was light (
IL); for the other half (group B) it was entrance size (IE). 

Post-treatment: colonies repeated the original choice to determine whether experience had altered their preferences.

Colonies were more likely to prefer the darker nest with the larger entry hole after they had made a point of rejecting lighter nests (group A). Meanwhile, the group B colonies that had rejected the larger entry became more likely to pick the lighter nest with a small hole. Once a colony had decided, ‘we don’t like nests that have large entry holes’, they continued to find entry hole size more important that other attributes. And the same was true for colonies that had explicitly rejected homes that were too bright. They decided that light level was the most important thing to look for in a new nest.

This suggests that circumstances can affect which attributes matter most to ants. Rather than being mere automatons, ants are capable of being influenced by their experiences. Which means they aren’t as different from us as we thought.


Sasaki T, & Pratt SC (2013). Ants learn to rely on more informative attributes during decision-making. Biology letters, 9 (6) PMID: 24196516

Wednesday, December 18, 2013

Just for fun: Insects with gears

You thought humans had invented interlocking gears. Tell that to the planthopper Issus coeleoptratus. Here's an extreme close-up of the insect's nether regions.

Planthopper_gears
Gears on an insect's leg, as viewed from below. Credit: Malcolm Burrows


A close-up of the planthopper's gears. Credit: Malcolm Burrows
A close-up of the planthopper’s gears. Credit: Malcolm Burrows


You can see the gears in action below:



To find out how they use these gears (hint, these insects are called planthoppers)  check out Ed Yong's post at Not Exactly Rocket Science.



Tuesday, December 17, 2013

Toward a male contraceptive

A safe, reversible, effective male contraceptive is almost the white whale of reproductive health. Thanks to work by Carl White from Monash University and his colleagues, there may soon be one. At least if you’re a mouse.
Over the years, there have been many efforts to create a male contraceptive. Most of these have used hormonal methods that focus on eliminating and/or debilitating sperm so that there can be no fertilization. Unfortunately, such strategies tend to have one or more serious drawbacks. They can be irreversible, they can increase the risk for birth defects in any future children, and they often have undesirable side effects on male sexuality.

White and his colleagues came at it from a different angle. They decided to block the movement of sperm through the vas deferens. This is a tube leading from the testis, where sperm is produced, to the urethra in the penis. When men have vasectomies, it is their vasa deferentia that are permanently severed. While vasectomies are usually highly effective, they're difficult or impossible to reverse.


Diagram of the ducts of the male reproductive system

The movement of sperm through the vas deferens isn’t a passive process, rather, the sperm are actively propelled by contractions within the vas deferens. Those contractions are caused by the activation of special receptor proteins embedded along the length of the vas deferens. In male mice bred without such receptors, there is no contraction of the vas deferens and the movement of sperm is stalled. This lack of sperm movement was 100% effective in preventing pregnancy.

To be clear, there’s nothing wrong with the sperm. They just can’t get to the business end of the penis. Sperm that was removed from these receptor-less mice were perfectly normal and capable of fertilizing mouse egg cells and making healthy baby mice. The adult male mice that were missing the receptors also seemed to be fine. Behaviorally, the altered mice were no different than the control mice. Their health was not compromised and nor was their appreciation for the lady mice.

While the mice used in this study were born with dysfunctional vas deferens receptors, there are also drugs that can block those receptors and they can be administered orally. I’m sure this will be much preferred over the currently being tested RISUG male contraceptive, since that one has to be injected directly into the penis.

Although White’s method sounds extremely promising, there is one potential problem. The same receptors that are in the vas deferens are also present in many other structures, including blood vessels and nerve cells. The mice didn’t appear to be harmed by the disabling of those receptors, but extreme caution should be used before starting any human trials.


White CW, Choong YT, Short JL, Exintaris B, Malone DT, Allen AM, Evans RJ, & Ventura S (2013). Male contraception via simultaneous knockout of α1A-adrenoceptors and P2X1-purinoceptors in mice. Proceedings of the National Academy of Sciences of the United States of America PMID: 24297884.



Monday, December 16, 2013

People use preventative care when they can get it


It’s obviously too soon to know how the Affordable Care Act will effect health care in the United States. However, researchers from the University of California, Davis and the University of Rochester have completed a study to see how having any type of health insurance affects individuals. In particular, the scientists were interested in whether people adopt riskier behavior and/or use more health services after acquiring insurance, one of the fears of people opposed to universal coverage.

Lead author Anthony Jerant and his colleagues analyzed data from nearly 80,000 adults over an eight year period. For each person, information on whether they had gained or lost insurance coverage, whether they had used any preventative care like vaccination or cancer screenings, and whether they had changed any health habits related to weight, smoking or use of seatbelts.

After gaining insurance, people were more likely to take advantage of preventative care. They were less likely to use such services after they had lost their health insurance. No surprises there. People are using services that are made available to them. However, people did not change their health behaviors, either for better or worse, when they gained or lost insurance. 

This is good news for insurance companies. People don’t suddenly start taking potentially expensive risks with their health just because they now have insurance. It’s also good for the patients that they recognize the value of preventative care and use it as much as they can. And by the way, this practice also saves money in the long run. 

As I said, these results do not include data from insurance changes under the Affordable Care Act. However, if one could extrapolate, this study certainly suggests that more people having access to health insurance would be a good thing for everyone.


Jerant A, Fiscella K, Tancredi DJ, & Franks P (2013). Health insurance is associated with preventive care but not personal health behaviors. Journal of the American Board of Family Medicine : JABFM, 26 (6), 759-67 PMID: 24204073.




Friday, December 13, 2013

DNA mutations might not be so random after all

File:Animal mitochondrion diagram en.svg

Some weird things are happening with our mitochondrial DNA (mtDNA). Mutations in mtDNA don't seem to be random, as David Samuels and his colleagues from Vanderbilt University School of Medicine found when they compared mtDNA obtained from ten different tissues from unrelated individuals. 

When we talk about our ‘genomes’ we usually mean the full complement of all our nuclear DNA divided into 23 pairs of chromosomes. However, our nuclei are not the only source of DNA in our cells. Mitochondria also have DNA (mtDNA).
The most likely explanation for why mitochondria have their own DNA is that far, far back in our evolutionary history, one cell engulfed an independently living bacterium and, rather than simply digesting that microbe, the larger cell harnessed the resources of the smaller cell. We all have these small power generators inside our cells, and they retain their own separate genomes from way back when they were free-living cells.

There are two interesting things about mtDNA. First, mtDNA comes exclusively from the female line. Not only is the human egg cell some 85,000 times larger than a human sperm, but the egg actively digests away everything in the sperm except DNA.  Therefore, the resulting baby begins life with a starter set of mitochondria that it inherited only from its mother.   

The second interesting thing, and the more important thing for our purposes, is that mtDNA is much smaller than the DNA that’s found in our nuclei. The human genome includes some 20,000 genes. In contrast, mtDNA only has 37 genes. As you can imagine, it’s a lot easier to determine the exact sequence of mtDNA than of a person’s full genome. This means that you can compare entire mtDNA genomes from person to person, or even from tissue to tissue within the same person. Thus, if a mutation turns up anywhere in mtDNA, you can find it.

There are two ways in which a person might end up with mutations in his mtDNA. First, he might have inherited the mutations from his mother. That is, those exact mutations might have already been present in a mitochondrion sitting in the egg cell before conception. In that case, cells from every organ or tissue in the body would have mitochondria with that same mutation.

Second, sometime after fertilization, a mutation could have occurred within a single mitochondrion. As that cell divided, all its daughter cells would share that mtDNA mutation, and more importantly, only its daughter cells would have that exact mutation. The mtDNA mutations would be tissue specific.

So which was it? Did the scientists find the same mutation in all ten types of tissue, indicating that the alteration in the DNA had occurred before fertilization, or did they find different mutations in each type of tissue?

Samuels and his colleagues found that the answer was: neither. When they compared mtDNA across tissues, they found that some tissues, particularly kidney and liver, shared the same specific mutations that were not found in the other tissues from the same individual. 

This is significant because our liver and kidneys originate from different embryonic tissues. The liver forms from the endoderm (inside layer of cells) whereas the kidneys are part of the mesoderm (middle layer), two tissue types that separated while the embryo was still a tiny ball of cells. If a mutation within a single mitochondria had been responsible for the changes in both the liver and kidney mitochondria, that mutation would have had to have occurred so early during embryogenesis that all the other bodily tissues would also have that mutation. But the other types of tissue did not have the same mutations. 

Even more perplexing, the same mutations were found in four different, unrelated individuals. That is, the exact same mutations that were found in the mtDNA of one person’s liver, but not the mtDNA of his skin cells were also found in other people’s livers but not skin cells.

Why would kidney and liver cells, which differentiated from each other in the first few weeks of embryogenesis, share the same mutation but kidney and skin cells, both derived from the mesoderm, not share the same mutation? And why would the same mutations show up in the same pattern in unrelated people?

The only explanation is that the same mutation occurred independently in the liver and kidney samples. 

Genetic mutations are supposed to be random. The chance of any particular change in the DNA sequence is small. It therefore falls to reason that it would be nearly impossible for an individual to have the same mutation occur independently in two different tissues. It would be even less likely for the same mutation to occur in the same tissues of two or more individuals. 

In other words, there must be some, as yet not understood mechanism that is altering mtDNA in prescribed ways. The researchers noticed that many of the changes occur in regions of the mtDNA that are involved in the replication of that DNA. It could be that the mitochondria with these mutations have some sort of advantage, or that these mutations are a common symptom of aging.

One thing’s for sure. We have a lot more to learn about mtDNA and about DNA mutation rates.


Samuels DC, Li C, Li B, Song Z, Torstenson E, Boyd Clay H, Rokas A, Thornton-Wells TA, Moore JH, Hughes TM, Hoffman RD, Haines JL, Murdock DG, Mortlock DP, & Williams SM (2013). Recurrent Tissue-Specific mtDNA Mutations Are Common in Humans. PLoS genetics, 9 (11) PMID: 24244193.



Thursday, December 12, 2013

Obesity affects taste receptors


Obesity is huge problem. Sorry. I’ll be good from now on. Anyway, one of the challenges for combatting obesity is to get people to eat less. Unfortunately, that’s made all the harder by decreased sensitivity to taste. According to Amanda Maliphol, Deborah Garth and Kathryn Medler from The State University of New York at Buffalo, diet-induced obesity can change one
’s taste perception.
To be fair, the researchers conducted their tests in mice rather than human volunteers. The mice were from a strain that easily develops obesity under the right conditions, and the researchers were happy to give them those conditions. Half the mice were fed high fat mouse chow, and the other half were given regular mouse chow. As expected, the mice grew at least 30% heavier on the high fat diet.

After 10 weeks, taste receptors were harvested from the mice. When a taste receptor responds to a taste, calcium channels open within that receptor. Thus, researchers can look for a calcium response to see whether the receptor had been activated. The obese mice had fewer taste cells responding to sweet taste, and the cells that did respond did so at a lower rate than in the control mice. The taste cells from obese mice also had a diminished response to bitter taste.

Whether one is satisfying a sweet-tooth or avoiding unpleasant foods, these results don’t bode well for reducing dietary intake. The authors are quick to point out that they can’t connect behavioral changes to these findings. However, it’s not hard to speculate that finding one’s food less sweet and/or less bitter does not make it easier to eat less of it.



Maliphol AB, Garth DJ, & Medler KF (2013). Diet-induced obesity reduces the responsiveness of the peripheral taste receptor cells. PloS one, 8 (11) PMID: 24236129.





Wednesday, December 11, 2013

Just for fun: Mosaicultures


This year, Montreal hosted the MosaĂŻcultures Internationales, a horticultural event extraordinaire. Unlike topiary, which mainly involves pruning, mosaiculture uses a variety of techniques, including painting and sculpture. You might think of the results as stationary Rose Parade floats.

It's too late to see the exhibit in person, but Alexander at My Virtual Garden has given us a virtual tour that you won't want to miss. 

Here's a sample. 

A 35 foot high Lady with cranes:

Lady and Cranes/ Chinese Myth


A 20 foot high shepherd:

The Old Man Who Planted Trees
Horses, Dog and Sheep

And an equally tall Mother Earth:

Mother Earth

Definitely check out My Virtual Garden for background information, plus lots more pictures.


Tuesday, December 10, 2013

Bacteria add ancient DNA to their genomes

When an organism dies, the DNA within its cells is released to the environment. Once there, the long DNA strands are quickly broken apart and degraded until almost all the fragments are shorter than 100 nucleotides. These DNA fragments are a source of food for bacteria, but apparently that’s not all the microbes can do with them. According to Søren Overballe-Petersen from the University of Copenhagen and his colleagues, these bits of DNA could be driving bacterial evolution.

The researchers found that the soil bacteria Acinetobacter baylyi weren’t just consuming small bits of DNA. The microbes were integrating the fragments into their own genomes via a process called 'transformation'. This was true even if the DNA was damaged with nicks or gaps, and even if it was as small as 20 base pairs long. 


Illustration of bacterial transformation. DNA from dead cells gets cut into fragments and exits the cell. The free floating DNA can then be picked up by competent cells.

Especially intriguing, the bacteria were able to incorporate DNA recovered from a 43,000 year old woolly mammoth. You might think that ancient DNA rarely becomes available to modern bacteria, but tons of the stuff is constantly being released from river sediments.

Not all types of bacteria can take up DNA in this manner. But for those that can, virtually any sequence of DNA on Earth is available for inclusion in that cell’s genome. After all, if you’re only talking about small snippets of DNA, chances are some animal has died and released that fragment within the past few thousand years.

Just to be clear, these small slivers of DNA are much too small to add new genes to the bacterial repertoire. Your average gene is at least ten times longer than the largest of these fragments. They can however, alter the bacterial genome, much like any other type of mutation would do. Thus, the incorporation of small bits of DNA, ancient or otherwise, may be yet another way that bacteria can evolve so rapidly.


Søren Overballe-Petersena, Klaus Harms, Ludovic A. A. Orlando, et al (2013). Bacterial natural transformation by highly fragmented and damaged DNA Proceedings of the National Academy of Sciences of the United States of America DOI: 10.1073/pnas.1315278110.



Monday, December 9, 2013

Temperature matters for mouse tumors

Animal models are enormously useful. Whether you’re studying concussions in flies or the progression of death in nematode worms, you can learn many things that are applicable to humans without actually harming any humans. For example, cancer therapies rely heavily on mouse models. There’s just one problem. The way the mice are housed may be affecting the data.

In most facilities, lab mice are kept at 2026°C (6879°F). This is very comfortable for the animal care technicians in their lab coats, but not so much for the mice, who prefer a balmy 30°C (86°F). This means that the mice are under a constant state of cold stress. Nobody thought much about this, since the mice seemed to be coping fine. However, as Kathleen Kokolus and her colleagues from the Roswell Park Cancer Institute found, coping and thriving are two different things, especially when it comes to cancer.

The researchers compared tumor formation and growth in mice housed at either 22°C or 30°C. After allowing the mice to acclimate to their maintenance temperature, the mice were injected with one of four types of tumor cells. Each type of tumor grew more rapidly in the cold mice than in the warm mice. The tumors also metastasized to the lungs more quickly at the colder temperature. When mice were allowed to move between cages set at different temperatures, healthy mice spent most of their time at 30°C, whereas tumor-bearing mice preferred 38°C, the warmest choice available to them.

Caption: Lab mice huddling in a cage.
Credit: Image courtesy of Kathleen Kokolus and Sandra Sexton.

Further tests implicated cytotoxic T cells (or CD8+ T cells) in delaying the growth and metastasis of tumors. Cytotoxic T cells are responsible for killing cells that have been damaged by infection or cancer, and there were more of these immune cells present at warmer temperatures. When the temperature comparison tests were rerun with immunocompromised mice that can’t make cytotoxic T cells, there was no difference in tumor growth between the warmer and colder mice.

The point of these experiments is not to suggest that human cancer patients should be kept warm (though obviously, people should be made as comfortable as possible). Rather, it’s to point out that unknown or unexpected variables can skew medical tests. This may explain the unfortunately common occurrence of experimentally promising drugs not living up to expectations in human clinical trials.


Kathleen M. Kokolus, Maegan L. Capitano, Chen-Ting Lee, Jason W.-L. Eng, Jeremy D. Waight, Bonnie L. Hylander, Sandra Sexton, Chi-Chen Hong, Christopher J. Gordon, Scott I. Abrams, & Elizabeth A. Repasky (2013). Baseline tumor growth and immune control in laboratory mice are significantly influenced by subthermoneutral housing temperature Proceedings of the National Academy of Sciences DOI: 10.1073/pnas.1304291110.