@Czarcasm
@Cawolf
How many cells does the organism have when this process occurs?
If the number of cells is small, isn't it theoretically possible that you could have someone with heterogenous alleles for an X-linked dominant disease show the recessive phenotype (if all the X chromosomes with the dominant allele were inactivated)?
From wikipedia:
It is understood that X-chromosome inactivation is a random process, occurring at about the time of gastrulation in the epiblast (cells that will give rise to the embryo). The maternal and paternal X chromosomes have an equal probability of inactivation. This would suggest that women would be expected to suffer from X-linked disorders approximately 50% as often as men[
citation needed] (because women have two X chromosomes, while men have only one); however, in actuality, the occurrence of these disorders in females is much lower than that. One explanation for this disparity is that over 25% of genes on the inactivated X chromosome remain expressed, thus providing women with added protection against defective genes coded by the X-chromosome. Some suggest that this disparity must be proof of preferential (non-random) inactivation. Preferential inactivation of the paternal X-chromosome occurs in both marsupials and in cell lineages that form the membranes surrounding the embryo,
[12] but does not occur in the vast majority of our cells.
The time period for X-chromosome inactivation explains this disparity. Inactivation occurs in the epiblast during gastrulation, which gives rise to the embryo.
[13] Inactivation occurs on a cellular level, resulting in a mosaic expression, in which patches of cells have an inactive maternal X-chromosome, while other patches have an inactive paternal X-chromosome. For example, a female heterozygous for haemophilia (an X-linked disease) would have about half of her liver cells functioning properly, which is typically enough to ensure normal blood clotting.
[14][15] Chance could result in significantly more disfunctional cells; however, such statistical extremes are unlikely. Genetic differences on the chromosome may also render one X-chromosome more likely to undergo inactivation. Also, if one X-chromosome has a mutation hindering its growth or rendering it non viable, cells which randomly inactivated that X will have a selective advantage over cells which randomly inactivated the normal allele. Thus, although inactivation is initially random, cells that inactivate a normal allele (leaving the mutated allele active) will eventually be overgrown and replaced by functionally normal cells in which nearly all have the same X-chromosome activated.
[14]
http://en.wikipedia.org/wiki/X-inactivation