Until a few years ago, the genetic variation of humans was understood only in terms of superficial characteristics, such as hair and skin colour. Today, thanks to the advent of cheap, fast genetic sequencing and DNA-microarray technologies, population geneticists can chart such variations in a more systematic way. Yet most experts agree that these studies are still in their infancy.

So it was with understandable incredulity that researchers received a plan by the UK Border Agency to use genetics to determine nationality — specifically, the origin of asylum-seekers claiming to be from war-torn Somalia. The agency's pilot programme, which began last month, aims to determine whether some 100 individuals really are Somali nationals by checking them for the individual DNA variants known as single nucleotide polymorphisms (SNPs) in mitochondrial DNA, on the Y chromosome and elsewhere in the genome. The scheme will also use isotopic ratios of elements found in hair and fingernails — which can vary depending on a person's diet or environment — to try to establish where the migrants previously lived.

The idea that genetic variability follows national boundaries is absurd.

The border agency says that the project has undergone scientific peer review, although it is difficult to say by whom: several geneticists contacted by Nature saw a preliminary proposal from the UK government in 2007, and warned that it was unlikely to work.

It is true that the recent development of large SNP databases have made it possible to determine the geographic origins of Europeans to within a few hundred kilometres (see Nature 456, 98–101; 2008). But comparable data on many human populations, especially in regions such as Africa, remain patchy at best, and it is unclear what data the border agency will use to establish the origins of these particular asylum-seekers.

On a more fundamental level, the idea that genetic variability follows man-made national boundaries is absurd. Cross-border migration is common throughout the world; Y-chromosome analysis can easily be thrown off by a distant male ancestor; and SNP-based identifications are inexact to say the least. As an example of this last point, individuals whose parents come from two geographic regions are often classed into a third region from which neither parent originated.

The use of isotopic analysis for identifying nationality is also unproven. Although it may be possible to use isotopic ratios to determine the region in which a person has recently lived, it cannot provide definitive evidence of their origins.

These problems seem to be ignored in the guidelines provided to border agents testing the asylum-seekers. Given the scientific credibility of DNA evidence, it is not difficult to imagine that these agents — who are presumably not geneticists — might place undue weight on results that are, at best, difficult to interpret and, at worst, spurious.

Migration organizations and geneticists alike have been vocal in their protests against the plan, and in response the UK government seems to have backpedalled. In a statement released earlier this week from the Home Office, which runs the border agency, the programme was described as only a proof-of-concept project that would not be used to make decisions about any asylum-seeker. But the government should cancel this scientifically dubious and politically sensitive programme outright. If it is allowed to continue, it could easily lead to a public backlash in the very populations that geneticists need to study to understand human origins and the genetic underpinnings of disease. Geneticists, and indeed all scientists, should decry the plan and make it clear that the science does not support it.