The Weismann barrier is the principle, proposed by August Weismann, that hereditary information moves only from genes to body cells, and never in reverse. In more precise terminology hereditary information moves only from germline cells to somatic cells (that is, soma to germline feedback is impossible). This does not refer to the central dogma of molecular biology which states that no sequential information can travel from protein to DNA or RNA.
The Weismann barrier is very important as it has implications for human gene therapy. If the Weismann barrier is permeable, then genetic treatments of somatic cells may actually result in a heritable change to the genome, possibly resulting in the genetic engineering of the human species rather than just individuals. It also has implications in our understanding of evolution as it would imply that species aren't nearly as separable genetically as we once thought. Furthermore it would open the door to the existence of certain Lamarckian concepts that previously had no supporting mechanism.
The use of this theory is sometimes referred to as Weismannism.
In the late 20th century there have been criticisms of an impermeable Weismann barrier. These criticisms are all centered around the activities of an enzyme called reverse transcriptase.
Evidence has begun to mount for horizontal gene transfer. Different species appear to be swapping genes through the activities of retroviruses. Retro-viruses are able to transfer genes between species because they reproduce by integrating their code into the genome of the host and they often move nearby code in the infected cell as well. Since these viruses use RNA as their genetic information they need to use reverse transcriptase to convert their code into DNA first. If the cell they infect is a germline cell then that integrated DNA can become part of the gene pool of that species.
Other evidence against Weismann's barrier is found in the immune system. A controversial theory of Edward J. Steele's suggests that endogenous retroviruses carry new versions of V genes from soma cells in the immune system to the germ line cells. This theory is expounded in his book Lamarck's signature. Steele observes that the immune system needs to be able to evolve fast to match the evolutionary pressure (as the infective agents evolve very fast). He also observes that there are plenty of endogenous retro-viruses in our genome and it seems likely that they have some purpose.
However, even if both of these possible exceptions turn out to be legitimate, the Weismann barrier loses only its absolute status, and may still remain impermeable in most cases. Without further examples, the penetration of the Weismann barrier is still very much an exception.
In plants, genetic changes in somatic lines can and do result in genetic changes in the germ lines, because the germ cells are produced by somatic cell lineages (vegetative meristems), which may be old enough (many years) to have accumulated multiple mutations since seed germination, some of them subject to natural selection.