X-inactivation (also called lyonization) is a process by which one of the copies of the X chromosome is inactivated in some female mammals. The inactive X chromosome is silenced by it being packaged in such a way that it has a transcriptionally inactive structure called heterochromatin. As nearly all female mammals have two X chromosomes, X-inactivation prevents them from having twice as many X chromosome gene products as males, who only possess a single copy of the X chromosome (see dosage compensation). The choice of which X chromosome will be inactivated is random in placental mammals such as humans, but once an X chromosome is inactivated it will remain inactive throughout the lifetime of the cell and its descendants in the organism. Unlike the random X-inactivation in placental mammals, inactivation in marsupials applies exclusively to the paternally derived X chromosome.

The paragraphs below have to do only with rodents and do not reflect XI in the majority of mammals. X-inactivation is part of the activation cycle of the X chromosome throughout the female life. The egg and the fertilized zygote initially use maternal transcripts, and the whole embryonic genome is silenced until zygotic genome activation. Thereafter, all mouse cells undergo an early, imprinted inactivation of the paternally-derived X chromosome in 4-8 cell stage embryos. The extraembryonic tissues (which give rise to the placenta and other tissues supporting the embryo) retain this early imprinted inactivation, and thus only the maternal X chromosome is active in these tissues.

The descendants of each cell which inactivated a particular X chromosome will also inactivate that same chromosome. This phenomenon, which can be observed in the coloration of tortoiseshell cats when females are heterozygous for the X-linked gene, should not be confused with mosaicism, which is a term that specifically refers to differences in the genotype of various cell populations in the same individual; X-inactivation, which is an epigenetic change that results in a different phenotype, is not a change at the genotypic level. For an individual cell or lineage the inactivation is therefore skewed or 'non-random', and this can give rise to mild symptoms in female 'carriers' of X-linked genetic disorders.

It is hypothesized that there is an autosomally-encoded 'blocking factor' which binds to the X chromosome and prevents its inactivation. The model postulates that there is a limiting blocking factor, so once the available blocking factor molecule binds to one X chromosome the remaining X chromosome(s) are not protected from inactivation. This model is supported by the existence of a single Xa in cells with many X chromosomes and by the existence of two active X chromosomes in cell lines with twice the normal number of autosomes.

The differentiation of phenotype in heterozygous females is furthered by the presence of X-inactivation skewing. Typically, each X-chromosome is silenced in half of the cells, but this process is skewed when preferential inactivation of a chromosome occurs. It is thought that skewing happens either by chance or by a physical characteristic of a chromosome that may cause it to be silenced more or less often, such as an unfavorable mutation.

The X-inactive specific transcript (Xist) gene encodes a large non-coding RNA that is responsible for mediating the specific silencing of the X chromosome from which it is transcribed. The inactive X chromosome is coated by Xist RNA, whereas the Xa is not (See Figure to the right). X chromosomes that lack the Xist gene cannot be inactivated. Artificially placing and expressing the Xist gene on another chromosome leads to silencing of that chromosome.

DNA packaged in heterochromatin, such as the Xi, is more condensed than DNA packaged in euchromatin, such as the Xa. The inactive X forms a discrete body within the nucleus called a Barr body. The Barr body is generally located on the periphery of the nucleus, is late replicating within the cell cycle, and, as it contains the Xi, contains heterochromatin modifications and the Xist RNA.

The existence of genes along the inactive X which are not silenced explains the defects in humans with abnormal numbers of the X chromosome, such as Turner syndrome (X0) or Klinefelter syndrome (XXY). Theoretically, X-inactivation should eliminate the differences in gene dosage between affected individuals and individuals with a normal chromosome complement. In affected individuals, however, X-inactivation is incomplete and the dosage of these non-silenced genes will differ as they escape X-inactivation, similar to an autosomal aneuploidy.

Besides, measuring the methylation (inactivation) status of the polymorphic human androgen receptor (HUMARA) located on X-chromosome is considered the most accurate method to assess clonality in female cancer biopsies. A great variety of tumors was tested by this method, some, such as renal cell carcinoma, found monoclonal while others (e.g. mesothelioma) were reported polyclonal.