However, there is no need for couples considering IVF to be unduly concerned. According to the study, there is only a 1 in 4000 chance of an IVF child having Beckwith-Wiedemann (BWS). So the risk is tiny, given that any child, conceived naturally or not, has roughly a 1 in 30 chance of having a major birth defect.

BWS is caused either by mutations in DNA or by errors in imprinting – the process in which chemical markers are added to DNA to switch off different genes in eggs and sperm. It can make certain parts of the body, such as the tongue, grow larger than normal and it increases the risk of certain cancers.

“But it’s not a catastrophic condition that leads to intellectual delay or major physical problems,” says Jane Halliday of the Murdoch Childrens Research Institute in Melbourne.

Most rigorous

Three papers published in 2003 suggested that IVF children might have an increased risk of BWS, but various flaws in the methods used have led to the reliability of these studies being questioned. The latest study is the most rigorous yet.

Halliday’s team analysed birth records from the state of Victoria between 1983 and 2003. Out of more than a million births, there were just 37 cases of BWS, a prevalence of 1 in 36,000. Of the 15,000 IVF children, however, four had BWS, of whom three were conceived by standard IVF and one by ICSI.

Comparisons of the 37 cases with aged-matched controls supported the finding that IVF children have a much higher risk. The results appear in The American Journal of Human Genetics (vol 75, p 526).

The study raises the issue of whether IVF babies should be screened for BWS during pregnancy. Halliday does not believe this is necessary. But fertility expert Mario Sousa of the University of Porto in Portugal thinks couples should at least be given the option. “I think it’s the honest way to do things,” he says.

In the future, it might be possible to screen embryos for imprinting errors before they are implanted in the womb, using pre-implantation genetic diagnosis, but at the moment such tests take too long to be practical.

Culture clash

Then there is the question of why IVF babies have a higher risk of this particular disorder. One possibility is that the process of IVF itself is to blame. It has been suggested that growing embryos in a culture medium can lead to imprinting errors.

If so, there might be an even higher risk of BWS with the new technique of blastocyst transfer, where the embryo is grown in culture for up to five days, rather than being transferred after two or three days. Blastocyst transfer seems to have a higher pregnancy rate.

Alpesh Doshi, an embryologist at University College Hospital London, does not think would-be parents will be deterred. “Patients will still be prepared to take that risk and have a higher chance of pregnancy.”

There is also growing evidence that the sperm or eggs of people with fertility problems are more likely to have mutations or imprinting errors. “The infertile patient has to consider that his own gametes might have some kind of higher risk of transmitting subtle defects,” says Sousa, who recently showed that the sperm of men with a low count have a higher rate of one kind of imprinting error.

Most studies of IVF suggest there is no increased risk of birth defects compared with natural conception. For ICSI, where the sperm is injected into the egg, the evidence is mixed.

A few studies have found a higher rate of malformations (91av print edition, 12 July 2003), and there have been several reported cases of ICSI children having another rare imprinting disorder called Angelman syndrome. ISCI also allows any genetic mutations responsible for a man’s infertility to be passed on to his sons.

Acne is a common and sometimes disfiguring complaint, affecting more than 80% of US adolescents. A number of factors are involved, including the bacterium and hormone levels.

But, until now, the importance of P. acne’s role was unknown. Holger Brüggemann, who sequenced the microbe with colleagues at the Göttingen Genomics Laboratory, Germany, says it was simply thought that if a large number of bacteria were present, it would trigger the inflammation and immune response associated with acne.

The new genomic data shows that the bacterium can produce proteins that actively cause acne. “P. acnes was regarded as a normal, harmless skin inhabitant – it wasn’t known that this bacterium has got a disease-causing potential,” says Brüggemann.

The team sequenced the 2.5 million bases in the genome of a P. acnes strain and identified 2333 genes, including some which code for enzymes that break down human skin.

“Sequencing the whole genome has revealed that the bacterium can actively degrade human skin tissue because of the massive presence of these enzymes, and also that there are specific immunogenic proteins which are present in this bacterium which trigger the immune response,” Brüggemann told 91av.

Contaminated blood

The fact that the microbe can be actively pathogenic raises the possibility of a potential public health threat from contamination of blood bank samples. Previous studies have found P. acnes in donated blood, introduced when skin flakes are dislodged during the processing of blood products or when an injection site is not properly sterilised.

Jochewed B. Werch, Chief of Transfusion Service, at the Ben Taub General Hospital in Texas has studied this problem, but believes that contamination is not likely to be a severe threat.

This is because the bacterium is extremely slow-growing, particularly in blood. “We cannot completely discount it, however,” she says, “because it has the potential.”

As well as highlighting risks, the P. acnes genome could also help develop new acne treatments. Severe acne is usually treated with common antibiotics, but many strains are becoming resistant to these. “With the genome sequence it’s now quite easy to generate specific drugs against this bacterium,” says Brüggemann. “That’s the next task.”

Mistaken identity

The genome also reveals a P. acnes gene mistakenly annotated as a human gene in the international Genbank database. This raised the possibility that if human sequence stored there is contaminated with bacterial sequences, other human sequence data may be similarly affected.

However, after investigating this, Julian Parkhill at the Wellcome Trust Sanger Institute, near Cambridge, UK, told 91av that the gene “has never been part of the human sequencing effort, it’s part of another effort to randomly sequence fragments of DNA from human tissue”.

He notes there is always a level of contamination in shotgun sequencing – the strategy employed for the GenBank sequence.

Journal reference: Science (vol 305, p 671)