Stem cell scientists are often irritated by the way people confuse their work with cloning, even though cloning plays no part in most ES cell research today. One reason for confusion is simply that both fields involve creating embryos.

Another may be an accident of timing: human ES cells were first cultured soon after the birth of Dolly, and commentators immediately pointed out the potential for combining the two discoveries. The term "therapeutic cloning" was coined to describe the creation of a cloned embryo as a source of ES cells; the embryo is destroyed in the process. In contrast, reproductive cloning would produce a baby from the cloned embryo.

Yet there is no denying that cloning is an important item on the stem cell research agenda, because it seems the best way to overcome a serious clinical problem with cell and organ transplantation: immune rejection. The immune system attacks any graft that is not genetically identical to the patient. Even a well-matched transplant requires lifelong treatment with immunosuppressive drugs, which have serious side effects, including increased susceptibility to infection and cancer.

Therapeutic cloning uses somatic cell nuclear transfer (SCNT), the technique that gave rise to Dolly: the nucleus of one of the patient's cells is transferred into a donated egg whose own nucleus has been removed. The egg is then stimulated to behave as if it has been fertilised, developing into an embryo that could be a source of ES cells with the same DNA as the patient. (Opponents of cloning point out that the same embryo could be implanted into a womb and grow into a baby.)

Unfortunately, SCNT is an inefficient process, in animals and people. The first scientifically credible account of human cloning came last year from Woo Suk Hwang and his colleagues at Seoul National University; they used 242 eggs to obtain 30 early embryos, from which they derived just one viable line of ES cells. South Korea has a culture of egg donation for research, which enabled the scientists to obtain good-quality eggs.

Indeed, even if therapeutic cloning can be made efficient, it is hard to see how enough human eggs could be made available to use the procedure in the clinic on a large scale (unless there is an unforeseen technical breakthrough).

In the more immediate future, however, scientists hope to use therapeutic cloning as a research tool that could give new insights into disease. While genetic disorders such as cystic fibrosis can be studied by deriving ES cells from embryos known to carry the single defective gene in question (see main article), this is not possible for diseases that result from multiple or unknown factors.

Last month Hwang's group in Korea announced the derivation of ES cell lines cloned from a range of patients suffering from inherited diseases or spinal cord injury. The efficiency of the process has improved, too: 185 donated human eggs yielded 31 cloned embryos and 11 ES cell lines. Lab tests confirmed that each cell line was immunologically compatible with the patient from whom it was derived.

Meanwhile other researchers are looking for alternative approaches to reducing immune rejection of stem cells. Some say even that the whole issue may have been exaggerated, because embryonic and foetal cells are intrinsically less immunogenic than adult cells--and they point out that neural transplants, for example, to treat Parkinson's disease, will benefit from the fact that the immune system is less active in the brain than elsewhere in the body.

One approach would be somehow to engineer the stem cells to make them less immunogenic or more compatible with the patient. A more drastic alternative would be to wipe out the patient's immune system and reconstruct it to match the transplanted cells. Some researchers have floated the idea of developing "universal donor cells" that would be compatible with everyone. But it is not clear whether any of these methods would work in practice.

Perhaps more achievable, though still an ambitious long-term project, is the idea of minimising rejection, rather than avoiding it altogether, by building up stem cell banks with many hundreds or thousands of cell lines representing as complete a spectrum of immune profiles as possible. Any patient in need of stem cells could then expect to receive a good if not a perfect genetic match.