Nature Podcast

This is a transcript of the 15th April edition of the weekly Nature Podcast. Audio files for the current show and archive episodes can be accessed from the Nature Podcast index page (, which also contains details on how to subscribe to the Nature Podcast for FREE, and has troubleshooting top-tips. Send us your feedback to


The Nature Podcast is sponsored by NPG Language Editing, a high quality English language editing service from the publishers of Nature.

End advertisement

Geoff Brumfiel: This week finding the missing genes behind complex traits.

Leonid Kruglyak: We're really excited when we could first do this and see all these genetic effects coming up just before us from were it was hidden from us.

Charlotte Stoddart: Working out why some cancer cells spread.

Elaine R. Mardis: Most oncologists will tell you that patients don't die from their primary disease; they die from their metastatic disease. So in that context we saw that it was very important at least continue this look at metastatic cancer

Geoff Brumfiel: And what happens when the lights go out, why power grids are prone to fail. This is the Nature Podcast. I'm Geoff Brumfiel.

Charlotte Stoddart: And I'm Charlotte Stoddart. When the human genome was mapped 10 years ago, scientists expected to find the genes underlying many diseases and traits, but they were disappointed. It turned out that most disease and traits like height and weight are much more complex, probably resulting from interactions between dozens, if not hundreds of genes. Genome wide association studies enable us to identify a small fraction of these genes, but most remain elusive. Well, Leonid Kruglyak and colleagues from Princeton University have finally found a way of uncovering these hidden genes, at least in yeast. Nature 464, 1039–1042 (15 April 2010)

Leonid Kruglyak: We've been trying to look at genetics of complex traits in yeast for quite a while and I've been involved in a lot of studies in humans and in other systems and its always been a little frustrating because you may be getting a small number of genes out and you can show statistically that there is a lot of additional complexity going on, but you can't actually get your, put your finger on it or see exactly what's going on and so we're very excited when we could first do this and see all these genetic effects coming up just before us from were it was hidden from us.

Charlotte Stoddart: So why is it so difficult to detect most of the genes that underlie these complex traits?

Leonid Kruglyak: I think basically the core issue is that because there's so many factors, the effect of any one factor individually are small and so any, sort of, statistically gaps will be really hard press to pick it up.

Charlotte Stoddart: But you have managed to find these missing genes, how have you overcome these problems?

Leonid Kruglyak: So we're working in a model system in yeast, we're not working in humans and that has a lot of advantages, but the main one that we've been able to take advantage is that we can have really enormous sample sizes, compared to what's possible in pretty much any other system, so because these organisms are you know, microscopic single celled organisms, we could grow easily billions of them in a flask.

Charlotte Stoddart: Okay, so you've got your billions of yeast in the lab, what do you do then to find these genes?

Leonid Kruglyak: Right, the trick is that basically these billions of cells, you can think of as, you know, a billion children from one family because they start with a pair of parent strains that we allow to mate and generate progeny and then they want to trap what regions of the genome contribute to trait variation that we see.

Charlotte Stoddart: And what traits are you looking at in your yeast?

Leonid Kruglyak: Primarily, what we've looked at in this study is resistance to drugs and different chemicals and the trick is basically that we first spread them out on plates that contain the drug whose effects we want to study as our trait and after a period of time that's long enough for the selection to take place, we take the cells that survive that treatment and we know the sequences both of the parent strains and what we want to know where in this group of survivors, they preferentially get DNA from one parents as opposed to the other.

Charlotte Stoddart: Right. So you expect 50-50 from mother and father's genome, but where you find a region of the genome where almost all of your cells have the father's genome, so you know that that part of the genome must be important for the particular trait you're looking at.

Leonid Kruglyak: Right. Exactly, so the father in that case must get a version of a gene in that region that increases resistance to the drug.

Charlotte Stoddart: So for many of the traits you said you did in yeast, you found lots of different genes underlying the variation, telling us that as we suspected many traits are genetically very complex, but can your study tell us anything about complex traits in humans?

Leonid Kruglyak: In this paper, we basically studied one single very large family and one of the debates about what's going on in human genetics is whether when you look at something like height at its inheritance within family, is that there is a fairly common set of genetic factors that's going to play at all in most individuals or there are lots of private genetic factors that effect what's going on in this in a family, but then when you look at a different family, it's a different set of such factors. And so that can really reach the level of complexity, so its possible that within any one family, you have you know, 20 factors, but every time you look at a new family, most of them will be unique an so that's how you get to these much larger numbers and that's something that we're starting to investigate in this system that we developed.

Charlotte Stoddart: It seemed to me like the more we learn about the genome, the more complicated the picture gets and we just realize that almost the more we know the less we understand.

Leonid Kruglyak: Yeah, I mean, I think we're getting clear or an exactly how complex these traits are, so I guess in that sense we're learning something, but in terms of actually being able to explain the genetics of common traits of common diseases, they're still pretty far away from having anything like a full picture.

Charlotte Stoddart: Is this all just about explaining and understanding the genetics of these complex traits or can we use then to cure help cure diseases in anyway for example.

Leonid Kruglyak: Right, I mean, I think they're at least two avenues people would like to think about exploring and going from this gene to medical applications, one is prediction in terms of diagnostics or risk assessment where you might choose to tell individuals different preventive strategies and so depending on their genetic makeup and improved diagnosis and so on and I think one of the things we're learning is that that's very hard. The other avenue that was identifying some of these individual genes even if their effects are minor, in terms of explaining variations in the populations is that they can be important players in the physiology of the disease and may lead to drug targets, regardless of how big a chunk of genetics they explain and I think that's something that a lot of our attention is focussed on.

Charlotte Stoddart: Leonid Kruglyak

Geoff Brumfiel: Now regular presenter Kerri Smith isn't here this week. She recently won the American Academy of Neurology Journalist of the Year award and she's picking up the prize in Toronto. But before she left, she caught up with cancer researcher, Elaine Mardis.

Kerri Smith: We've reported on a fair, few cancer genome studies on the podcast, the usual process of sequencing a cancer genome involves taking a block of cells from a particular tumour and sequencing the cumulative genome from all of those cells, then a team might sequence a control sample, a piece of normal, healthy tissue from the same patient. The idea is to find out how cancers differ from normal cells and how this might effect how they grow and importantly how they spread. This week a team led by Elaine Mardis of Washington University in St. Louis, Missouri report four genome sequences from a patient with breast cancer, the primary tumour, a healthy sample from the patient's blood and two more as Elaine explained to me. Nature 464, 999–1005 (15 April 2010)

Elaine R. Mardis: So in this study, we were particularly fortunate to also have from this same patient a second tumour, metastases that had formed about 8 months after her primary diagnosis of breast cancer and this metastatic tumour was actually presenting in the frontal lobe of her brain and was removed by a neurosurgery process. The fourth sample in our sample court, as we call it locally, was a so called xenograft or human and mouse tumour. The primary tumour from this patient was sampled, while the tumour is still in the patient's body and then implanted into an immunocompromised mouse, so that we can study it and one of the fundamental questions that has not yet been addressed about the xenograft model, these human and mouse tumours is exactly how much genomic change happens once that tumour is removed from the human body and placed into the mouse body.

Kerri Smith: So you've got these four samples then, the human tumour in the mouse and then the three from the different parts of this patient's body and is that because you thought that the mutations might be different in each place or?

Elaine R. Mardis: Well that was really the question, there were 16 genes that were impacted by mutation in the primary disease, but they were actually present in a very small percentage of the tumour cells in the primary tumour, conversely in both the metastatic tumour and in the xenograft, these 16 mutations are shown to be present in a very large number of cells or effectively every cell in that tumour.

Kerri Smith: Could it be that the mutations that are present in these distant sites, these metastases, be the cause of the cancer actually being able to move?

Elaine R. Mardis: Right, perhaps what is happening is that a small collection of cells are actually leaving the primary tumour and able to propagate successfully in another distant body site, in this case it was the brain.

Kerri Smith: So these genes are basically the cancer's ticket to ride.

Elaine R. Mardis: They could be yeah.

Kerri Smith: What does that mean for clinically for treatment if cancers in different locations just have a very small number of different mutations?

Elaine R. Mardis: Well, it's a little hard to extrapolate what it means for clinically, I mean, most oncologists will tell you that patients don't die from their primary disease; they die from their metastatic disease. So, in that context, we felt that it was very important to at least continue this look at metastatic cancer. The downside is that it's difficult to infer too much from just a single instance of primary to metastatic transition.

Kerri Smith: So this, sort of, tells you more about the evolution of a cancer over time than it does necessarily for treatment.

Elaine R. Mardis: It does indeed and that was the really interesting glimpse that we got from comparing the primary metastatic as well as the xenograft tumour because the surprising outcome from our paper really is that what happened to the genomes of a metastatic tumour and of the xenograft tumour turned out to be very, very similar and that's a little unexpected, because the xenograft tumour the one that implanted into the mouth was taken from the patient before she went through any chemotherapy and any radiation therapy. Conversely, the metastatic tumour essentially survived, if you will, both the combination of chemotherapy and radiation therapy and yet the genomic results between the metastatic tumour and the xenograft turn out to be largely very similar. This makes us wonder about whether the xenograft model actually turns out to be a very good way of studying the metastatic process.

Kerri Smith: One last question for you Elaine, do you think this is what personalized cancer genomics and treatment might have to look like in the future in order for, you know, this genomic information to be useful.

Elaine R. Mardis: I really do, actually, and I often get to ask the question well would you have your genome sequenced and my answer to that has always absolutely not, unless I had cancer. And I think and I predict in many of my talks that cancer patients will actually be the first patients to begin to benefit from personalized genomics because each cancer seems to be quite different and more to the point of your question there are now targeted therapy, so that if you a mutation or an alteration in a specific gene, there may be a targeted therapy or moving forward into the future a combination of targeted therapies that could be combined in a cocktail essentially, to address each and every of the mutations in your cancer genome.

Geoff Brumfiel: Kerri Smith there talking to Elaine Mardis. You're listening to the Nature Podcast. Now for the best of the rest from Nature, here are the headlines.

Charlotte Stoddart: The boom in discovering exoplanets is set to accelerate with a new technique reported in Nature this week. It's hard to spot planets that orbit very close to their stars because the light from the planet is washed out by scattered light from the star itself. Now a team from the Jet Propulsion Lab at Cal Tech have a device rather thrilling called a vortex coronagraph that acts a starlight filter, a bit like noise cancelling headphones, but for telescopes. The team pointed the scope at three HR8799 recently discovered exoplanets 130 light years away, to prove they could pick them up. Nature 464, 1018–1020 (15 April 2010)

Geoff Brumfiel: Last year, scientists successfully replaced 40 mitochondrial DNA in monkeys and now it seems the same gene therapy could be used in humans. Mitochondria are the power generators of our cells and they have their own DNA. Some times disease causing mutations occur in this DNA and are then passed down directly from mother to child. A team at Newcastle University has found they can reduce the amount of mitochondrial DNA that's inherited by transferring nuclei between human female eggs. By letting the eggs develop into embryos in a test tube, they found the recipient eggs only carried over 2% of mitochondrial DNA, well below the diseased threshold. Nature advance online publication 14 April 2010

Charlotte Stoddart: What was the earth like 3.5 billion years ago, when life was still in its infancy? We know that back in the Archaean era, all life was single celled and found in the oceans and we had thought those oceans were hot around 55 to 85 degrees Celsius, but recent studies hinted a much cooler sea. Scientists infer ocean temperatures from chemical fingerprints in ancient rocks. When a team from Yale analyzed the oxygen isotope composition of phosphates in sediments in Southern Africa, they concluded that the Archaean ocean was no more than 35 degrees, supporting another recent study that limited the temperature to 40. Nature 464, 1029–1032 (15 April 2010)

Geoff Brumfiel: Now a bit of poetry for the pod. For a want of a nail, the shoe was lost; for want of a shoe the horse was lost; for want of a horse the rider was lost; for want of a rider, the battle was lost; for want of a battle that kingdom was lost; and all for want of a horse shoe nail. That proverb dates back to the 14th Century here in England and for those who studied network theory, its still relevant today. This week Eugene Stanley at Boston University and his colleagues have published work describing how small failures in interconnected networks can have catastrophic consequences. I called Eugene to learn more. Nature 464, 1025–1028 (15 April 2010)

H. Eugene Stanley: For want of a nail, the shoe was lost and so forth. It's a wonderful and insightful proverb because it makes clear that you can indeed have a catastrophic cascade, really bad news, what could be more catastrophic than to lose the entire kingdom. However, it's so rare we don't run around you know, being careful about nails. We don't do that because it's very, very, very rare. So the damage of our work, we showed clearly if not dramatically that this was not the case, when networks are coupled. When networks are coupled, it's not at all rare that you'd have a catastrophic cascade.

Geoff Brumfiel: It seems to me that it's sort of obvious that interconnected networks would behave differently than single networks, why hadn't people thought to look at this before or had they?

H. Eugene Stanley: First of all, they have. People have at first thought about this in networking in their works but the main thing that we've added is the actual demonstration by mathematics, if you will that you can have a very abrupt collapse of a system of coupled networks.

Geoff Brumfiel: So take me through in 2003, there was a massive power black out in Italy which I gather was, sort of, one of the things you checked your results against in this paper. Tell me what happened with this power black out.

H. Eugene Stanley: On September 28th, the power went out in Italy, so that's the fundamental fact, power was out. Now what caused the power to go out, the naïve view would be that some transformer burned because a transformer was critical to the functioning of a network, nothing else could function, but that simply is not the case. The power companies built in redundancy, a transformer can burn out, and you still have power; you don't lose power for the whole country of Britain, simply because one transformer burns out because things burn out all the time, so the reason power went out is not that the reason is the power in one transformer so to speak was found to control a computer node as part of the computer network and therefore that computer node became non-functional, that means all the other computer nodes connected to it became non-functional and therefore all the power grid nodes elsewhere could not function; and then there's more and more powered grid node through a now more and more computer grid node through a now back and forth and back and forth and therefore the experience of the person in the street was that suddenly there was no power all over Italy.

Geoff Brumfiel: And this really is the counterintuitive result of your paper, right. I mean, normally in networks we think the more connections there are in a network, the more robust it is, the more routes there are, if there's a failure of the one nodes, but what you've found here is that the more nodes you have interconnected between two networks, the more vulnerable the networks become, isn't that right.

H. Eugene Stanley: Exactly right. But the key thing is two networks, if you'd had only network, for example, if particularly the computer grid do not rely on the power grid, each one relied on battery then it would reasonably robust, but because it depended on another network, which was vulnerable it was not robust, then things happen.

Geoff Brumfiel: Just thinking about it, I mean, the real power grid might be even more interconnected than just with the internet. I could easily imagine you know, the power stations depend on the rail networks, I suppose, to get fuel to the plant. Does this teach us anything about interconnected networks that would allow us to ensure against these failures.

H. Eugene Stanley: Yes and no. I mean just the key statement is that the tightly coupled infrastructures are extremely vulnerable. And that's what we need to be aware of because each decade the world is characterized by more tightly coupled infrastructures and as we go along encouraging these infrastructures, it sounds good, but as we do this we have be cognizant of the fact that they become vulnerable, and what we do about that is a good question, I mean, that's what people who know about how to handle vulnerability, but I assume it just to be prepared that something bad can happen.

Geoff Brumfiel: Eugene Stanley. Now beyond Nature's shores, science also happens and news editor Mark Peplow has joined us in the studio to enlighten us on science news of the past few days. Hi Mark.

Mark Peplow: Hello Geoff.

Geoff Brumfiel: So first up this week, I guess we have a story about children who truly don't see colour, they don't see racial boundaries, is that right?

Mark Peplow: Yeah that's right. These are a group of children with a pretty rare neuro developmental disorder called William's Syndrome. Now they tend to be sort of overly friendly and they don't fear strangers at all and a study comparing a group of 20 children with William's syndrome with 20 children who don't have that disorder has actually picked up some really interested differences with how they see gender and how they see race.

Geoff Brumfiel: So what are the differences, so let's start with race then? What are the differences there and how they perceive race?

Mark Peplow: When you actually have a group of children who you know, standard white European descendant children, the experiment that they did was to actually tell them a little story in this case, they tell a story of two little boys, one of them is a kind little boy, he saw a kitten fall into a lake and he picked the kitten up to save it from drowning, so which is the kind little boy and they showed them a picture of a little black boy and a little white boy and they get the kids to pick one which is the good one that saved the kitten. Now in the kids without this disorder, basically you tend to see some kind of racial bias, they will attribute negative attributes to the black kid, but in the William's syndrome kids, there's no sense of that at all, they're really racially colour blind. Now, interestingly this is different from their reaction to when you ask them to do gender stereotyping, for example, if you have drawings of a little boy and a little girl and you ask them who is playing with the doll, then everybody picks the girl. So what the researchers behind this study are saying is that they've managed to show that you can biologically dissociate if you like, these two different forms of stereotype

Geoff Brumfiel: And did they have any idea how they're dissociated or why?

Mark Peplow: In terms of the actual mechanism of how this works, William's syndrome people have abnormal activity in a brain structure called amygdala and that's involved in responding to things like to social threats and triggering unconscious negative emotional responses. And what they're suggesting is that their experiments shows that this is strong evidence that its social fear that leads to racial stereotyping because its that's part of the brain, the medulla, responsible for social fear that's actually not working properly in these kids.

Geoff Brumfiel: Ha, fascinating. So its sounds like an interesting theory, but the ideas of race are so complicated, I mean there's so intertwined with social cues and all sorts of other things, how can be they sure this is really what's going on.

Mark Peplow: Well, the researchers who are based in Germany do say that they need to do studies on larger samples and with broader age groups to actually confirm this and while some the neuroscientists that Nature spoke to and things that really changes the picture to really think what you actually mean by stereotyping and others would suggest that may be possibility that kids with William syndromes may have actually had different experiences of members of other racial groups. It may be that their parents for whatever reason are treating them slightly different from normal kids and those talking to them about races are not at all are in different ways than they would do with children with out this disorder. So, that's definitely something that needs to be close up off to absolutely confirm this theory.

Geoff Brumfiel: Okay. So moving on from issues of race and childhood, next up here we have a story about sushi and specifically whale sushi.

Mark Peplow: Yeah that's right. This is a fantastic tale of scientific sleuthing, a group of scientists at Oregon State University, who have basically investigated where certain restaurants actually got their sashimi from, both of the restaurants, one in California, one is South Korea, were advertising this whale meat for sale. The key question in this investigation is where did that whale meat come from

Geoff Brumfiel: And so what did they find, where is it from?

Mark Peplow: Basically they did a genetic analysis on whale meat samples taken from these restaurants and they've compared it with samples taken from whale meat purchased in Japan and they're shown basically that whale meat that's being sold in a particular Santa Monica sushi restaurant and in south Korea both came from whales that were caught in Japan as part of the Japanese whale catch for supposedly scientific research.

Geoff Brumfiel: So I mean, this must really deal a blow to the Japanese. They claim they're doing these whale catches for scientific research and finding its way onto commercial market. Is this, I guess, its politically probably quiet sensitive, is it really surprising then.

Mark Peplow: I don't think they're overly surprising. I mean, people know that whale meat is sold from Japan to different parts of the world and the key question is of course, whether it's strictly allowed. For example is south Korea, whale meat is legal there, as long as it caught accidentally during other fishing operations, proving that the whale meat that they were serving there actually came from Japan, caught deliberately, actually means that it shouldn't actually be being sold there and indeed the office actually notified the metropolitan police about that. I think the key thing is that going forward, in June the International Whaling Commission is going to consider a new proposal on whaling, which would allow Japan and Norway and Iceland to hunt whales while actually reducing their corteges. The key thing about the genetic analyses that's being done here is that that proves there was a mechanism that you can actually monitor those corteges you can actually monitor where whale meat is coming from and where its going.

Geoff Brumfiel: And last up on the list, there is some good news for stem cell scientists this week.

Mark Peplow: Yeah, two of the most widely used embryonic stem cell lines appear to be perhaps just weeks away from eligible for US Federal funding and while some stem cells scientists are celebrating this, they're also chaffing that the really year long process that its taken to actually get these stem cell lines ready for approval for funding.

Geoff Brumfiel: So where does this leave stem cell research in the US now? I mean, it was on hold for years with the Bush administration. There wasn't any federal money going, is there enough, is it going to get enough lines to really make stem cell scientists happy?

Mark Peplow: Here's the very potted back plot, you remember that back in March 2009, Obama overturned the Bush era limits on federal funding to research embryonic stem cells basically both should say that you could only get federal funding for research on stem cell lines that have been derived before 2001 and that meant that there was only really 21 lines in circulation. Obama said no, you can apply for approval on a whole host of lines as long as you go for the proper ethical procedures to make sure that consensus has been obtained from donors and things like that, potentially openings up 400 to a 1000 new stem cell lines. So there's been a lot of excitement and indeed since then the NIH, the National Institute of Health has approved 51 new lines for federal funding. The problem is that only one of the original bush era lines alliance has been approved since then and that's largely because of problems with paperwork over sort of, the original consents from the donors that were obtained and the trouble is because people have been restricted to those bush lines for so long most of the work has been done on those and you need to keep doing work on those to compare their properties with new lines that are coming through under the Obama administration. And so delays from actually getting approval to get funding to do new research in these lines has really held things up in the stakes over the last year.

Geoff Brumfiel: Okay, well great, well thanks very much Mike. All that and more at

Charlotte Stoddart: Join us next time, for a post-mortem of the Copenhagen climate change meeting and resurrecting the Dead Sea. I'm Charlotte Stoddart

Geoff Brumfiel: And I'm Geoff Brumfiel. Keep track of your nails.


Maintaining NPG's standards of excellence Nature Research language editing helps you deliver the best possible manuscripts so your research results stand out. Upload your manuscript at today.

End advertisement