Donate

If you would like to support our clinical research you can make a tax-deductable donation to Ped IMID.

Navigation
Tuesday
Jan122010

The role of sex in evolution

Sex is a powerful force for evolution. On the face of it, sex seems like an absurdly complicated way to reproduce. Prokaryotic organisms, bacteria and archea, have a much faster a simpler system, where the cell simply duplicates its DNA and splits in half into two identical daughter cells. The entire process, called mitosis, only takes 20 minutes. This means that under ideal circumstances a single bacterium can divide to produce 8 offspring in the first hour. In the second hour that single precursor cell could form 64 offspring, after 6 hours a single cell could form over 200,000 daughter cells. This asexual reproduction is so efficient that it only operates at capacity for very short durations, as exponential growth of a single cell could use up the resources of an entire planet within days. Typically a bacterium ticks over slowly by scavenging what resources are available, only to explode into exponential asexual growth when new resources become available and a race to exploit them occurs.

Compare this to the elaborate, time-consuming and often bizarre process of eukaryotic sex, which multicellular organisms from plants to fungi to animals use to reproduce. Sex (and the accompanying mate selection) is one of the most difficult and dangerous parts of an individual’s life, and even passionate advocates of the activity find it difficult to explain. Yet through an evolutionary lens, sex provides very concrete advantages. The best illustration of the advantages of sex come from yeast mating, as these simple organisms are capable of both asexual and sexual reproduction.

Simple sex

Yeast can be thought of as being halfway between simple bacteria and complex multicellular organisms like humans. In terms of lifestyle and behaviour, yeast operate like bacteria – single celled organisms capable of an independent existence through the use of resources in their direct environment. Inside the cell, however, yeast are clearly eukaryotic organisms, with the same basic machinery for cell division, metabolism and survival as plants and animals. It is therefore convenient to think of yeast as essentially human-like cells, trapped in an early bacterial-like lifestyle. This is an oversimplification of course: bacteria, yeast and humans are all highly evolved organisms and none have remained static in evolutionary time, but it is a useful oversimplification.

So how do yeast reproduce? Asexually, like the bacteria they share a lifestyle with? Or sexually, like the multicellular organisms they are genetically closest to? The answer is both. When yeast are in a rich nutrient environment they reproduce asexually like bacteria. A single cell undergoes mitosis, duplicating its DNA and then splitting into two daughter cells, each identical to the parental cell. This gives the yeast all the advantages of bacterial reproduction – very simple rapid reproduction to win the race for abundant resources. The parental cell was successful in the environment, so the identical daughter cells should be equally successful and proliferate likewise.

However as noted above, exponential growth can never continue unabated, sooner rather than later resources become limiting or some other factor stresses the survival of the yeast. At this point yeast have a trick available that bacteria do not – sex. Instead of undergoing dormancy, the yeast mate.

In the best understood system, that of Saccharomyces cerevisiae, there are two sexes of yeast, a and a, controlled by a single gene. Mating is very simple, the a cells release a chemical called ‘a factor’ and produce a receptor that causes them to migrate towards the chemical ‘a factor’. By contrast, the a cells release a chemical called ‘a factor’ and produce a receptor that causes them to migrate towards the chemical ‘a factor’. The two yeast cells, one a and one a, attract each other and fuse into a single cell. This cell now has two different copies of the yeast genome, one from each parent.

The a-a fused yeast cell can now undergo a complicated cellular division process called meiosis. Unlike mitosis, where the cell duplicates its genome and divides in two, meiosis involves duplicating the genome and dividing in four. This is possible because the a-a fused yeast cell has two copies of the genome to start with, so duplication gives four copies, one for each of the four daughter cells that result.

The important difference between mitosis and meiosis is the splicing of two different genomes to form unique combinations. Mitosis just duplicates the existing genome. Meiosis starts with two different genomes, and during the duplication processes these genomes are jumbled up together, creating new combinations of old characteristics. This means that all four daughter cells at the end are unique and different from the original parental cells.

The advantage conferred by sex is very straight forward – the parental cells were not dealing well with the environment they were in, since yeast mating occurs only under stress. Therefore why reproduce more cells that cannot cope with the environment? Instead the yeast takes a life-or-death gamble that a combination of genetic information from another cell will produce offspring better able to deal with the environment. In a simple scenario there would be two yeast strains, one able to deal with acidity and one able to digest complex carbohydrates. A change in environment to a high acidity environment where the only resources available are complex carbohydrates will stress both parental strains. However, by sex there is a chance that one of the daughter cells will inherit the acid resistance of one parent and the ability to digest complex carbohydrates from the other parent. Other daughter cells will not be so lucky and will die, but that one daughter cell with the chance combination of two necessary characteristics will be able to divide asexually and rapidly reap the rewards of a new resource.

In one final complication, yeast can change sex. A single gene makes yeast either a or a, so after mating and meiosis the four daughter cells include two a cells and two a cells. If a single a cell is successful in the new environment, asexual reproduction creates exact copies, so all progeny will be a cells. This would create an obvious problem if a new environmental stress requires another round of mating, so yeast carry spare “silent” copies of a and a genes and use these backup copies to flip from one sex to another, to make sure a population is always a mixture of a and a yeast.

Saturday
Dec192009

The things they don't teach you about being a scientist

One of the frustrating issues in a science career is the limited extent to which each career stage prepares you for the next. An undergraduate degree in science will typically focus on teaching established science theories and testing them via examination. The research proportion is limited and shrinking due to budget constraints.

Then you finish undergraduate studies and start a PhD, and the ability to learn established theories and sit an exam is completely useless. Instead you need to completely reorientate yourself to research skills, both practical (in terms of benchwork) and theoretical (in terms of experimental design and analysis). Exactly where your PhD mentor expects you to pick up these new skills is a mystery, as there are no lectures or classes to teach it. Throughout your PhD and postdoc you know that you are going to be judged solely on your research output. Do experiments and publish, do experiments and publish, anything else is irrelevant.

So you finish your postdoc with a lot of research experience and a handful of publications and manage to land a faculty position. You are now an independent principle investigator and all the skills you have learned to date are redundant. No time for benchwork anymore, you need to master a new set of skills within a year or fail miserably and end with a whimper. Having only written one or two short fellowship applications at the end of your PhD, you now need to master the major project grant. A detailed and elaborate research proposal which needs to be tailored towards the language and politics of the particular granting body (information which is never given on the website of course), your grant has to compete with successful investigators who have been operating in the field for decades.

While you wait for a year for the grant results to come back, your startup grant seems to disappear - quick, learn the skills of an accountant! So far you've only spent money in the lab, now you need to know the complete salary costs (including taxation status, social security contributions and yearly increments), equipment depreciation costs, which items should go on which budget (international staff on the VIB budget can gain expat taxation status, but international students on the KUL budget are exempt from social security), the cost threshold for requiring multiple quotes, how to negotiate with reps for good prices, and so much more. When you have mastered this you realise that you wasted far too much money on furniture when the university has a hidden basement full of free cast-offs and that expensive piece of equipment you bought already exists unused in a laboratory two floors down.

Of course, while you are becoming a grant writer / accountant, research needs to occur, so you'll need staff. You are a complete unknown, so no high power post-docs coming with their own fellowship. You didn't teach undergraduate classes last year, so good luck in snapping up a student able to attract a scholarship. You place a few adds in Nature and for the next six months you get ten applications a day from India and China. How to judge them? Hiring decisions are a science in themselves, then labour contract law is a mine-field. Nevertheless, with a few bumps along the road you somehow manage to put together a surprisingly talented and hardworking team. You already knew from personal experience that a lab can be an emotional boiler-room, now you need to manage that or manage the consequences. You need to understand every staff member as an individual, what makes them tick, how to keep them happy and productive, the best way to redirect them when they go off-course. Skills that can take a lifelime to learn about your partner you need to pickup within a few months about six strangers from six different cultures. Plus you'll need to leave your computer enough to spot trouble brewing in the early stages. The small things matter, the person irritated about someone else casually borrowing pipettes and not returning them happens to have a habit of writing directly on glass bottles. And let's face it, scientists are not exactly trained in emotional intelligence.

Think that you can do research now? Equipment, check. Reagents, check. Grant money, check. Staff, check. Stir the pot and research comes out? Hah! You would be breaking a surprising number of regional, national and international laws. You'll need a liquor license for that ethanol to clean benches, a permit to use sedatives on mice, ethics clearance of course but also an animal use license. Biosafety permits, equipment certification, occupational health and safety monitoring, a fire-warden. The most frustrating part is that there is no check-list to work down, you only find out about a requirement when you think you are there and you hit a brick-wall.

Then there are the unpredictables, that sap away your time until you are ready to scream. Your immunology department is the only one in the world without a flow cytometry core unit. The research assistant you hired to look after the mouse colony turns out to be afraid of mice. Your contract unexpectedly stipulates that you become fluent in Flemish within three years. That assay you used to do in your sleep simply doesn't work in Belgium. Your post-doc falls in a legal loophole that makes them ineligible for fellowships designed for both locals and foreigners. The SPF mouse house didn't tell staff to set up breeders inside a hood and all your imported mouse strains are contaminated. Your weeks of slaving over an FWO grant are wasted because you didn't know that the FWO does not have anonymous peer review and requires you to submit your own reviewers. You find out that your start-up grant also has to cover your own salary and you over-hired in the first year. You have a hundred meetings, departmental politics, collaborations to foster and suddenly a year has gone by and you didn't even manage to finish off that project that was nearly ready for publication at the end of your post-doc.

Of course, I could be externalising. Perhaps I just missed the training session.

Sunday
Dec062009

The inefficient consequences of evolution

Vertebrates are unique in developing an immune system capable of anticipating pathogens that are yet to evolve. Birds and mammals have taken this "adaptive" immune system to the pinnacle, with T cells and B cells using a randomised form of genomic engineering. The advantage of a system based on randomised generation is striking - by making every T cell and B cell unique it becomes exceptionally difficult for pathogens to "out-evolve" their hosts. Regardless of how a pathogen will change, pre-existing T cells and B cells will be capable of recognising the new modified pathogen. The importance of the adaptive immune system to humans is evident in the fatal consequences of its absence, such as patients with end-stage AIDS or primary immunodeficiencies caused by genetic mutations. These benefits greatly outweigh the cost of the adaptive immune system in resources used and the threat of autoimmune disease.

But does the adaptive immune system make vertebrates more healthy? There is no obvious evidence that it does. In a key essay on the topic, Hedrick argues that vertebrates do not appear to have a lower pathogen-induced mortality rate than invertebrates. Instead, he argues that the development of the adaptive immune system provided only a short-term benefit, with pathogens rapidly being specialised to vertebrate hosts. The result is an immunological arms race, with each side incrementally ratcheting up the armaments. Vertebrates are essentially impervious to non-specialised pathogens unless rendered immunodeficient, but the additional mortality from specialised pathogens is probably equivalent to the invertebrate state.

This still-controversial hypothesis high-lights an important aspect of evolution by natural selection. It has highly inefficient consequences. Natural selection takes place at the level of the individual and evolution takes place at the level of the species. Most importantly, natural selection only occurs in the present. An individual who has an advantage for even a single generation will be over-represented in the next generation. A species that has an advantage for a single generation will be able to exploit more resources for reproduction. The long-term consequences - that each species will waste more resources in an ever more expensive battle - is irrelevant.

The evolutionary arms-race between host and pathogen is one incredibly important example. A more illustrative example of the patent futility of this arms-race comes from Sir David Attenborough, one of the leading science communicators of all time. In Life in the Undergrowth, he films two species of harvest ants living in the desert. Each population needs to collect seeds to survive, however the number of seeds produced in the desert is so low that there is fierce inter-species competition. One species of ant is diurnal, the other nocturnal, and each is capable of collecting the entire daily seed dispersal. In order to survive, every second night the nocturnal ants spend an evening carrying rocks to cover the entry hole of the diurnal ants. The diurnal ants can't collect seeds the next day as they need to spend a day clearing the rocks from the entrance. This gives the nocturnal ants a night to harvest the uncollected seeds. The following day the diurnal ants are able to collect every seed and that night the nocturnal ants spend carrying rocks. Two species end up literally carrying rocks backwards and forwards every second day.

The elegance of evolution is the beauty of such specialised behaviour, but the consequences are gross inefficiency in resource use. If each species simply spent alternative cycles conserving resources both species could survive with a higher population density than currently exists. But neither species can be the first to stop the wasteful use of resources, as that would give a fatal advantage to the other, and so they are trapped together in a cycle of carrying stones. The battles of night ants vs day ants and of hosts vs pathogens illustrate the bizarre, elaborate and ofttimes perverse consequences of evolution by natural selection

Friday
Nov202009

Science is not a family-friendly career

There is a brief article in this week's Nature entitled "Tenure or family?"

Marriage and childbirth are what stop most female US graduate students from becoming tenured researchers, according to a report by Washington DC think tank the Center for American Progress (CAP) and the University of California, Berkeley, School of Law. Staying Competitive: Patching America's Leaky Pipeline in the Sciences found that married mothers with a PhD are 35% less likely to enter a tenure-track position in the sciences than married fathers with PhDs, according to a National Science Foundation survey. And they are 27% less likely than their male counterparts to get tenure after securing a tenure-track post. The report advises universities and funding agencies to create family-friendly policies, including six weeks of paid maternity leave and a week of paid parental leave.


Obviously there is an enormous problem in career progression for women in science. A 35% reduction at the tenure-track stage and a 27% further reduction at the tenure stage - women get whittled out of the academic career pathway. This article kind of misses the point though. Marriage and children are not what stops women progressing in science. Extra maternity leave is not going to help if it puts women further behind the publication scramble. To put it bluntly, in my opinion this is the real problem:

1. A career in science is horribly unfriendly to a balanced life. There is no security or safety, every step of the way 90% of people are going to jump or be pushed. Everyone is smart at the top, that isn't enough, you also have to be lucky and obsessively determined. Most tenure-track professors don't even take weekends or holidays - they can't afford to be left behind.

2. Society still has structural sexism built in. Yes, women are now free to pursue any career they want, in addition to their previous workload. If it was purely child-rearing that was a problem the blockade would be in all scientists who have children. Instead the burden falls disproportionally on women scientists who have children, because on average they still end up doing more of the work than men. Consider the recommendations of the report: six weeks of paid maternity leave and a week of paid parental leave. Even if the recommendation is passed, women will be expected to do six times more child-rearing work than men.

These problems are much harder than simply paid parental leave, although obviously that would be a positive contribution. Instead we need to tackle the two fundamental issues. The science career needs to be made more family friendly, or at least not a horrific all-consuming ordeal. We can't continue with the same massive bottle-necks in careers or with a system where every person works themselves to death to stay in the game one more round. Competitive peer review has grown into a destructive monster that chews people up and spits them out. Secondly men need to pull their own weight rather than expecting women to sacrifice their time to make up for a thoughtless spouse.

Wednesday
Nov112009

A time-line for diabetes research

6th century BCE – The first known diagnosis of diabetes was made in India. Doctors called the condition medhumeha, meaning "sweet urine disease", and tested for it by seeing whether ants were attracted to the sweetness of the urine.

1st century CE – Diabetes was diagnosed by the ancient Greeks. Aretaeus of Cappadocia named the condition διαβήτης (diabētēs), meaning "one that straddles", referring to the copious production of urine. It was later called diabetes mellitus, "copious production of honey urine", again referring to the sweetness of the urine. Unlike the Indian doctors, Greek doctors tested this directly by drinking a urine sample. At the time a diagnosis of diabetes was a death sentence: "life (with diabetes) is short, disgusting and painful" (Aretaeus of Cappadocia).

It is probably that the ancient Egyptians and early Chinese cultures also independently discovered diabetes.

10th century CE - Avicenna of Persia provided the first detailed description of diabetes (diagnosed through "abnormal appetite and the collapse of sexual functions" as well as the "sweet taste of diabetic urine"). He also provided the first (partially) effective treatment, using a mixture of lupine, trigonella and zedoary seed.

1889 – Joseph von Mering and Oskar Minkowski in Germany developed the first animal model of diabetes using dogs, discovering the role of the pancreas.

1921 - Federick Banting and Charles Best in Canada first cured canine diabetes by purification and injection of canine insulin.

1922 - For the first time diabetes stopped being a death sentence. In 1922 Federick Banting and Charles Best treated the first human patient with bovine insulin. Notably they decided to make their patent available globally without charge.

1922-1980 - Treatment of patients with animal insulin or human insulin extracted from cadavers. Substantial life extension but also significant side-effects.

1955 - Determination of the protein sequence of insulin by Federick Sanger in the United Kingdom.

1980 - First commercial production of recombinant human insulin, by Genentech.

Today there is no cure for diabetes, but when treated it only results in an average loss of 10 years (the same as smoking).

Sunday
Nov082009

Polling on science in America

Interesting figures released by Pew on the opinion of the American public on science and scientists.

Effect of science on society

Mostly positive84%
Mostly negative6%

Contribute "a lot" to society's well-being

Military84%
Scientists70%
Clergy40%
Journalists38%
Business executives21%

US scientific achievements are the best in the world
American public agreement17%
American scientists agreement49%

Major problems for science, as identified by scientists:
The public does not know very much about science85%
Public expects solutions to problems too quickly49%
New media oversimplify scientific findings48%
Lack of funding for basic research46%

Differences of opinion between scientists and the public*:
Agreement with theory of evolution87% vs 32%
Agreement with theory of climate change84% vs 49%
Support for use of animals in research93% vs 52%
Support for compulsory vaccination of children82% vs 69%
Support for embryonic stem cell research93% vs 58%

* Presumably these would be even more striking if broken down by scientific discipline.

Scientific training changes ideology

US PublicUS scientists
Liberal20%52%
Moderate38%35%
Conservative37%9%

Friday
Oct232009

The ethics of biobanking

The University of Leuven hosted two lectures on biobanking today, one by Hainaut from the International Agency for Research on Cancer and the other by Juhl from the biobanking company Indivumed.

Biobanking is a tricky ethical area, with little consensus and vague law. Who owns the material taken from a patient? The patient? The hospital? The surgeon? If someone wants to use the material, what is the default position? Should the patient have to provide consent or is consent assumed unless the patient opts out? Does the patient even have the right to opt out at a latter time point? Hainaut made the case that there is a moral duty on every person to allow access to their biological samples for the good of humanity. His example was that a excised breast cancer not only belongs to that woman, but also to all other women who may develop breast cancer in the future.

This is an attractive argument but has flaws. If the information generated goes into the public sphere, such that new treatments can be developed and accessed, it may be reasonable to use the moral argument, in the same way that organ donation as the default option can be argued on moral grounds. However, to me this argument is flawed if the information generated does not go into the public sphere. If the information is not published (a secretive researcher or company keeping back information for potential future uses) or if it is published with restrictions on use (ie, patented) that information is not open to all of humanity. Isn't it unethical for a biobank to appeal to the moral duty to all of humanity unless legal restrictions are placed on the biobank to ensure that the proceeds of the bank are available to all of humanity? Doesn't informed consent require donors to be told the status of information generated from their samples?

Unfortunately, Hainaut was not able to answer this question when asked, as Juhl (CEO of a biobanking company that only publishes a fraction of the data it generates) jumped in with a rant about for-profit vs not-for-profit. His contention was that every person acts through the personal profit motive, so that whether the biobank made a profit or not didn't matter. His position is that only private companies have the money to put forward to do the research, and they deserve a profit for the research they do. Perhaps, but irrelevant to the ethical question. If the research outcomes are utilitarian then the utilitarian argument should be put to prospective donors - such as DeCode offering all future drugs free of charge to Icelandic people in exchange for access to the medical records and genome of the Icelandic people. Material can be collected for a utilitarian motive using utilitarian appeals, or for a moral motive using moral appeals. What is unethical is to use a moral appeal to collect material destined for a utilitarian purpose.

Hopefully we will see future legislation reflect the ethical considerations of biobanking in more a more thoughtful manner than was presented today. Donations made by the public for the public good should be legally bound to this use. It is illegal for a charity to accept a monetary donation, keep 90% of the money for personal use and spend 10% on charitable works. Likewise it should be illegal for a biobank that accepts material presented as a public donation to only release 10% of the data produced by the donation, and keep 90% to itself.

Monday
Oct192009

Infectious cancer

It has long been known that the several causes of cancer are infectious. Typically a virus contains a number of oncogenes to enhance its own proliferation, and in an infection gone wrong (for both virus and host) a viral oncogene is incorporated into the host DNA, creating an uncontrollable tumour cell. One of the best examples of this is human papillomavirus (HPV), a virus which infects most sexually active adults and is responsible for nearly every case of cervical cancer worldwide (which is why all girls should be vaccinated before they become sexually active).

However these cases are not "infectious cancers", they are infectious diseases which are capable of causing cancer. True infectious cancers, where a cancer cell from one individual takes up residency in a second individual and grows into a new cancer, were unknown until recently. With the publication of a new study in PNAS we now have three examples of truly infectious cancers.

1. In the most recent study, researchers in Japan documented the tragic case of a 28 year old Japanese woman who gave birth to a healthy baby but within two months had been diagnosed with acute lymphoblastic leukemia and died. At 11 months of age the child also become ill and was diagnosed with acute lymphoblastic leukemia. Genetic analysis of the tumour cells in the baby demonstrated that the tumour cells were not from the child herself, but rather maternal leukemia cells that had crossed the placenta during pregnancy or childbirth and had taken up residency in their new host. With this information, retrospective analysis indicates that this is probably not a one-off event, and at least 17 other cases of mother-to-child transmission of cancer have probably occurred.

2. In addition to mother-to-child transmission of cancer, cancer can spread from one identical twin to another. Identical (mono-zygotic) twins have identical immune systems, preventing rejection of "transplanted" cells, unlike non-identical (di-zygotic) twins. Thus a tumour which develops before birth in one identical twin can be transferred in utero to the other identical twin, where it can grow without being rejected. In one improbable but highly informative case, a set of triplets were born where two babies were identical and the third was non-identical. A tumour had arisen in one of the identical twins in utero and had passed to both other foetuses, but had been rejected by the non-identical foetus and accepted by the identical foetus. Of course, with the advent of medical transplantation, transmission of infectious cancers is now no longer limited to the uterus. Transplantation of an organ containing a cancer into a new host can allow the original cancer to grow and spread, as transplantation patients are immunosuppressed to prevent rejection. There is also a single case of a cancer being transmitted from a surgeon who cut his hand during surgery to a patient who was not immunosuppressed.

3. In a medical mystery well known to Australians, the population of Tasmanian Devils has been crashing as a fatal facial tumour has been spreading across the population. The way the fatal tumours have spread steadily across Tasmania and sparing Devils on smaller islands first suggested a new infectious disease that causes cancer, similar to HPV in humans. However a suprising study demonstrated that the cancer was directly spreading from one Devil to the next after having spontaneously developed in a single individual. These scrappy little monsters attack each other on first sight, biting each other's faces. The cancer resides in the salivary glands and gets transmitted by facial bites to the new Devil. Unfortunately for Tasmanian Devils, a genetic bottleneck left all Devils so genetically similar that they are, for immunological purposes, all identical twins. This means that the cancer cells transmitted from one Devil to another through biting are able to grow and kill Devil after Devil. The cancer from a single individual has already killed 50% of all Devils, and it is possible that we will have to wait until the cancer burns out by killing all potential hosts before reintroducing the Devil from the protected island populations. As unlikely as this seems, another similar spread occurs in dogs, where a cancer that arose in a single individual wolf is being spread through sexual transmission from dog to dog around the world. This example also illustrates the point made about cancers being "immortal" - the original cancer event may have occured up to 2500 years ago, with the tumour moving from host to host for thousands of years without dying out.

Saturday
Oct032009

When you eat matters

A very interesting study has just been published in the journal Obesity. The work, by Arble and colleagues in the Turek laboratory, fed mice high-fat food either during the day or at night. The surprising result was that mice fed during the day put on 20% more weight than mice fed at night. In both cases the mice had unlimited access to food yet both groups of mice ate the same amount, so there was no difference in net calories. Instead, what this result suggests is that the body deals with calories differently at different points of the diurnal cycle. During the active phase (night for mice) calories are shifted into burn mode, while during the resting phase (daytime for mice) calories are stored with greater efficiency.

If this result can be translated into humans it would suggest that large meals should be concentrated in the active phase of the day, breakfasts and lunches, and that evening or night meals should be restricted. An interesting proposal is that the American evening-biased eating rhythm compared to the European lunch-biased eating rhythm is partly responsible for the obesity problem in America. Of course it could only ever be a fraction of the problem, as many other correlates with obesity are well recognised. For example, a study by Pickett and colleages has demonstrated that countries with higher income inequality have higher calorific intake and obesity, and another study by Bassett and colleagues points out that Belgians burn 62 extra Calories per day by walking and cycling, compared to a poor 20 Calories per day by Americans.

The other important aspect of this study is that it contributes to the growing body of evidence dispelling the simplistic "obesity = too many calories and not enough exercise" formula. As published by the Segal laboratory, the majority of difference in body mass index (BMI) is due to genetics (64%). Being overweight does not mean that an individual is making worse eating or exercising decisions than a healthy range individual - the majority of the difference in weight just comes down to the fact that different genetics leads to different metabolisms.

Tuesday
Sep222009

Nature attacks peer review

In the latest issue of Nature, the journal has published a rather unfair attack on peer-review. Peer review is the process that most journals use to assess the merit of individual papers - submissions are judged by editorial staff, then sent to scientists working in the field for peer review, then the reports by these scientific peers are judged by the editorial staff to determine whether they warrant publication. While it is the standard today, there has been a lot of resistance to peer review in the past, as the editorial staff of journals exercised their power of selection. Notably Nature, founded in 1869, only moved towards peer review 100 years later, under the direction of John Maddox. Other journals, such as PNAS, are only now scrapping peer review bypasses.

There are certainly problems with the journal submission process, but typically these involve too little peer review, rather than too much. A journal such as Nature typically rejects the majority of papers without review and for those papers reviewed there are only two to three reviewers per paper. Scientists put a lot of effort into reviewing, but as it is an unpaid and unrequited favour, it is not the highest level priority. Even after review, the editorial staff have enormous power to accept or decline the advice of peer review, Nature once famously publishing a paper falsly reporting to show effects of homeopathy. This editorial decision tends to be a combination of ranking the news splash effect (Nature and Science compete for citations in the big newspapers), the "boys club" effect (no longer all male, but certainly the big names have an easier pathway to acceptance) and editorial "gut feeling".

To justify the editorial over-ride, defects in peer review are commonly cited. In this latest editorial piece, Nature presents the results of an unpublished study presented at a conference, reporting that the results show a bias of peer review towards positive results. This may be so, but does the cited study actually show that? What the study did was submit two papers, one with positive results and one with negative results, to two journals, and analyse the peer review results. The results showed that peer reviews at one journal (Journal of Bone and Joint Surgery) had a minor reduction in ranking the negative results paper, while the second journal (Clinical Orthopedics and Related Research) showed no significant difference. Hardly a damming inditement of peer-review.

What are the methodological flaws that could account for the minor differences observed at one out of two journals?

* Different reviewers. Even picking 100 reviewers for each paper does not cancel out this effect unless reviewers were carefully stratified to ensure random distribution.

* The quality of the two papers may have been different. The author of the study tried to make them as identical as possible, but different results need to be presented differently. As the study is unpublished we only have the author's opinion that the two studies were of equal quality.

* Positive and negative results can have very different "impacts". Most journals explicitly request a review which takes into account both scientific validity and scientific impact. Negative results generally have lower impact and hence would get lower review scores, as explicitly requested by the journals. To remove this effect the papers should have been submitted to a journal such as PLOS One, which requests a review only on scientific quality.

* Positive and negative results require different statistical standards. A positive result uses simple statistics to show that the two groups were different. A negative result requires more complex statistics and can only state that the two results were not different above a certain level. A negative result can never exclude that a positive result exists with a smaller effect than would be picked up by the study design.

Certainly the most obvious sign of "positive bias" evidenced by this article is the decision by Nature to write an editorial and broadcast a podcast on a minor unpublished study that denigrates peer reviewers and hence elevates editorial staff. Would they have written a similar editorial on an unpublished presentation showing no sign of bias by peer reviewers? The minor impact observed in one out of two journals tested (with all the caveats above) did not warrant Nature to fill its editorial with phrases such as "dirty", "biased", "more negative and critical" and "biased, subjective people". The worst bias of all is the accusation that peer reviewers from the second study only showed no statistical bias because "these reviewers guessed they were part of an experiment". Surely Nature should have been able to spot that subjective reporting, dimissing negative results and elevating positive results are the very definiton of positive result bias!