American Journal of Medical Genetics 129A:136 –143 (2004)
Inheritance of Most X-Linked Traits Is Not Dominant
or Recessive, Just X-Linked
William B. Dobyns,1* Allison Filauro,2 Brett N. Tomson,2 April S. Chan,2 Allen W. Ho,2 Nicholas T. Ting,2
Jan C. Oosterwijk,3 and Carole Ober1
1
Department of Human Genetics, The University of Chicago, Chicago, Illinois
2
The College, The University of Chicago, Chicago, Illinois
3
The Department of Clinical Genetics, Groningen University Hospital, Groningen, The Netherlands
The existence of X-linked disorders in humans has
been recognized for many centuries, based on
lessons in religious texts and observations of
specific human families (e.g., color blindness or
Daltonism). Our modern concepts of Mendelian
(including X-linked) inheritance originated just
after the turn of the last century. Early concepts of
dominance and recessiveness were first used in
conjunction with autosomal traits, and then appli-
ed to ‘‘sex’’-linked traits to distinguish X-linked
recessive and X-linked dominant inheritance. The
former was defined as vertical transmission in
which carrier women pass the disorder to affected
sons, while the latter was defined as vertical
transmission in which daughters of affected males
are always affected, transmitting the disorder to
offspring of both sexes. However, many X-linked
disorders such as adrenoleukodystrophy, fragile
X syndrome, and ornithine transcarbamylase defi-
ciency do not fit these rules. We reviewed the
literature on 32 X-linked disorders and recorded
information on penetrance and expressivity in
both sexes. As expected, penetrance and an index
of severity of the phenotype (defined in our
Methods) were both high in males, while the
severity index was low in females. Contrary to
standard presentations of X-linked inheritance,
penetrance was highly variable in females. Our
analysis classified penetrance as high in 28% of the
disorders studied, intermediate in 31%, and low in
40%. The high proportion of X-linked disorders
with intermediate penetrance is difficult to recon-
cile with standard definitions of X-linked reces-
sive and dominant inheritance. They do not
capture the extraordinarily variable expressivity
of X-linked disorders or take into account the
multiple mechanisms that can result in disease
expression in females, which include cell autono-
mous expression, skewed X-inactivation, clonal
expansion, and somatic mosaicism. We recom-
mend that use of the terms X-linked recessive
and dominant be discontinued, and that all
such disorders be simply described as following
‘‘X-linked’’ inheritance. ß 2004 Wiley-Liss, Inc.
*Correspondence to: William B. Dobyns, M.D., The University
of Chicago, Department of Human Genetics, Room 319 CLSC, 920
E, 58th Street, Chicago, IL 60637.
E-mail: wbd@genetics.bsd.uchicago.edu
Received 12 September 2003; Accepted 2 January 2004
DOI 10.1002/ajmg.a.30123
ß 2004 Wiley-Liss, Inc.
KEY WORDS: dominant; recessive; X chromo-
some; X-inactivation; X-linked
inheritance
INTRODUCTION
The ‘‘rules’’ for X-linked dominant and recessive inheritance
were set down by Morgan et al. [1915, 1922] more than 80 years
ago, although these were foreshadowed by ancient texts such
as the Talmud and by early reports of families segregating
X-linked traits, such as the pedigree pattern of colorblindness
(a.k.a. Daltonism) that was probably first pointed out by the
Swiss ophthalmologist Horner [Horner, 1876] in a local Canton
record (see [Thompson, 1986] for a biographical sketch of
Horner). The rules were adapted for human genetics during
the 1920s and 30s, and have changed little since then, despite
dramatic advances in our understanding of the molecular basis
of genetic traits. The rules relating to X and Y chromosome
segregation have a firm statistical basis and need not be
reconsidered. However, the literature and our personal ex-
perience with human X-linked disorders suggest that the
currently accepted rules related to mode of inheritance fail to
account completely for the wide variation seen in practice and
need to be reconsidered.
In this study, we revisit the origins of our concepts of
X-linked inheritance in Drosphila and their application to
human disorders, review currently accepted ‘‘rules’’ as set
down in standard Genetics texts, and compare expectations
based on these rules with actual observations of penetrance
and expressivity for 32 X-linked disorders. We show that
several accepted rules for X-linked inheritance conflict with
these observations, discuss the likely basis for these miscon-
ceptions, and attempt to resolve the conflicts by proposing a
revised set of rules. This problem was presented to an under-
graduate class in Human Genetics at The University of
Chicago in the winter quarter of 2002 (BioSci 21200), and the
class did much of this work. Some of the sources used in our
review have passed from the memory of modern science, but
still have much to offer. We have therefore taken the liberty of
citing several of them at length when they offer particularly
useful insights.
We will show that the concepts of X-linked recessive and
dominant inheritance, as currently used, are based on (1) a
system of dosage compensation for genes located on the X
chromosome that is entirely different from humans, (2) more
rigid definitions of the terms ‘‘dominant’’ and ‘‘recessive’’ than
were originally intended, and (3) superficial use of the vast
body of experience available regarding human X-linked
disorders. The terms X-linked dominant and X-linked reces-
sive work reasonably well in experimental settings using
Drosophila, but have never worked well in humans. We pro-
pose that both terms be dropped entirely, and that disorders
resulting from mutations of genes located on the X chromosome
be described simply as ‘‘X-linked.’’
 X-Linked Inheritance 137

Ancient Texts
Experience with X-linked diseases extends far back in
history. The Talmud shows clearly that ancient Jewish
scholars recognized the existence of disorders transmitted
through women. Jewish tradition calls for all newborn males to
be circumcised. Infant boys with X-linked hemophilia may die
from bleeding after circumcision. The Talmud teaches that the
death of either two or three infant boys after circumcision
represents a pattern or ‘‘chazakah,’’ and that subsequent sons
should not be circumcised. This pattern was recognized for sons
of an individual woman and sons born to sisters. For example,
one rabbi (R. Rebbi) argued that if a woman circumcised two of
her sons and they died, she should not circumcise her third son,
while another (R. Shimon Ben Gamliel) taught that she should
circumcise the third but not the fourth [Babylonian Talmud].
The latter rabbi also cited the case of four sisters, three of whom
circumcised their sons who all died. He advised the fourth
sister not to circumcise her son. Thus, the family history of
sisters was used to establish a legal pattern.

Dawn of Modern Genetics

The field of genetics as it exists today began with the
recognition of chromosomes as the hereditary material and the
rediscovery of Mendel’s work at the turn of the last century
[Benson, 2001]. Within a few short years, the founders of
modern genetics discovered independent segregation and re-
cognized dominant, recessive, and sex-linked traits. The first
rules for heredity including the biological basis for linkage and
sex inheritance were set down in Morgan’s principal concep-
tual work and now classic text ‘‘The Mechanism of Mendelian
Heredity’’ [Morgan et al., 1915, 1922], a book also known for
publishing the first linkage map of Drosophila (reproduced in
Fig. 1). The origins of our modern concepts of dominant,
recessive, and X-linked (originally sex linked) inheritance are
found here, so we began our efforts to reconcile historical
concepts of X-linked inheritance with modern knowledge of
molecular and medical genetics by reviewing these concepts as
originally presented.
Dominance and recessiveness. Conklin [1915] origin-
ally defined dominance as the ‘‘development in offspring of
certain characters of one parent while contrasting characters Fig. 1. Reproduction of the frontspiece from the 1922 revision of
of the other parent remain undeveloped.’’ Morgan et al. [1915, Morgan’s text [Morgan et al., 1922] showing the four linkage groups. Traits
1922] are never exact in their definition of the terms on the Drosophila X chromosome are listed as Group I.
‘‘dominant’’ and ‘‘recessive,’’ but rather cite examples of how
the terms were used. This first appears when they introduce
the inheritance of pairs of traits, then called factors or only in homozygotes are recessive. The degree of dominance
characters. ‘‘The mutant stock called vestigial is so character- was also not considered important. ‘‘Whether a character is
ized because it has only small vestiges of the wings. If a fly with completely dominant or not appears to be a matter of no special
vestigial wings is mated to the wild type with long wings, the significance. In fact, the failure of many characteristics to show
offspring will have long wings. If these hybrid flies of the first complete dominance raises a doubt as to whether there is such
generation (the first filial generation, or F1) are mated to each a condition as complete dominance’’ [Morgan et al., 1915,
other, their offspring (F2) will be of two sorts: some will have 1922].
long wings and others will have vestigial wings. There will be Sex-linked inheritance. By the time the Morgan et al.
three times as many flies with long wings as flies with vestigial book was written, biologists already understood the two types
wings. This is the Mendelian ratio of 3:1 that appears when a of sex inheritance [Morgan et al., 1915, 1922]. In the
single pair of characters is involved.’’ They go on to note that Drosophila type (XX-XY), which includes mammals and thus
factors for this trait must be located on a single pair of chromo- humans, females are homozygous for a sex factor while males
somes, and that the fertilized egg derived from a mother with are heterozygous. In the Abraxus type (WZ, WW), the pattern
the factor for long wings (wild type) and a father with the factor is reversed. In both types, heterozygous individuals mate
for vestigial wings (mutant) will carry both factors. Their only with homozygous individuals, resulting in equal numbers of
definition of the terms dominant and recessive follows. ‘‘Since heterozygous and homozygous individuals in each succeeding
this egg with both factors present produces a fly with long generation. Morgan and his contemporaries had identified
wings, the vestigial character is said to be recessive to the long; many new traits in Drosophila. Their studies soon showed that
or conversely the long is said to be dominant to the vestigial these characters were inherited in four groups, later shown to
character.’’ correspond to the four Drosophila chromosomes, and that
In general, traits expressed in both homozygotes and heter- traits in separate groups assorted independently. Within each
ozygotes are described as dominant, while traits expressed of these groups, segregation of traits deviated significantly
138 Dobyns et al.
from the expected Mendelian ratios (Fig. 1), observations that
supported the concept of linkage or gametic coupling as had
then only recently been reported in the sweet pea. They found
that one of these groups (their Group 1) was sex linked. Their
definition was simple: ‘‘certain factors follow the distribution of
the X chromosome and are therefore supposed to be contained
in them. These factors are said to be sex linked’’ [Morgan et al.,
1915, 1922].
Sex-linked dominant and recessive inheritance. The
rules for sex-linked inheritance are again illustrated by
example rather than stated explicitly. The first example given
for sex-linked inheritance is the white eye character. ‘‘If a
white-eyed male is bred to a red-eyed female (wild type), the
sons and daughters (F1) have red eyes. If these are inbred,
the offspring (F2) are 3 reds to 1 white, but the white-eyed flies
are all males.’’ In this example, red eye is dominant to white
eye. Males with the white eye trait have white eyes as the white
eye trait was inherited from the mother, while the Y chromo-
some from the father brings no dominant factor.
Morgan et al. [1915, 1922] cite many examples of sex-linked
characters in Drosophila (Group I in Fig. 1) and other species.
Most were recessive. ‘‘In the case of all the factor-pairs, except
abnormal abdomen and bar, the normal allelomorph dom-
inates. Therefore, the females will appear normal for all
characters except abnormal and bar, which are dominant.’’
However, their concept of dominance and recessiveness allow-
ed more flexibility than we generally permit now, as a single
character could be recessive in one situation but dominant in
another as they explain for white eyes. ‘‘For instance, in flies the
factor for white eyes seems to produce no effect when white is
bred to red. The F1 reds are indistinguishable from pure reds.
But by weakening the red by adding recessive factors other
than white, the influence of white can be demonstrated.’’
So, Morgan et al. would likely not have considered most human
X-linked disorders rigidly dominant or recessive.
Dosage Compensation in Drosophila and Man
Morgan et al. [1915, 1922] based their rules for sex-linked
inheritance primarily on their observations in Drosophila, and
then extended these to other species with the XX/XY or
Drosophila system of sex determination, including humans.
They were unaware that the sex determining mechanisms in
insects and humans are very different. In Drosophila, sex is
determined based on the number of X chromosomes in any
given cell [Parkhurst et al., 1990; Parkhurst and Meneely,
1994]. Cells with two X chromosomes are female, while cells
with only one are male. Genes on the Y chromosome are
expressed only later during sperm formation. Dosage compen-
sation for genes on the X chromosome occurs by transcribing
them more rapidly in single-X containing cells, producing as
much gene product as the two X chromosomes in a female cell
[Keyes et al., 1992; Parkhurst and Meneely, 1994; Kelley et al.,
1997].
In humans, sex is determined based on expression of a
signaling pathway initiated by the SRY gene on the Y chromo-
some [Sinclair et al., 1990]. Dosage compensation occurs by
inactivating most genes on the X chromosome in one of the two
X homologues in cells containing two X chromosomes. Thus, X0
individuals are infertile males in Drosophila, and infertile
females (45,X) in humans. These differences in dosage com-
pensation result in significant differences in gene expression.
In Drosophila females with null mutations of an X-linked gene,
100% of cells will have 50% expression of the gene product. In
human females with null mutations of genes on X that are
subject to X inactivation, 50% of cells will have 100% expres-
sion of the gene product and 50% will have 0% expression on
average (Fig. 2). Thus, a cellular phenotype would be apparent
in approximately half the cells of human females with random
Fig. 2. Diagram emphasizing the different cellular phenotypes that
result from heterozygous mutations of X-linked genes based on the
Drosophila and human forms of X chromosome dosage compensation. In
Drosophila, all heterozygous female cells will have half of gene/protein
function (shaded pink or gray). In humans, half of heterozygous female cells
will have normal gene/protein function (shaded red or black), while the other
half will have minimal or no function (shaded white). [Color figure can be
viewed in the online issue, which is available at www.interscience.wiley.
com.]
X inactivation. These differences are substantial, and could
well lead to differences in expression of the phenotype at the
level of the organism.
Hermann Werner Siemens and Sex-Linked
Inheritance in Humans
The concept of sex-linked inheritance was soon applied to
human genetics, with common disorders such as hemophilia
and some genetic eye diseases such as colorblindness recog-
nized as fitting the sex-linked recessive pattern described by
Morgan and others, in which males were affected and females
unaffected carriers. An important early report of sex-linked
dominant inheritance in humans is the detailed description by
Siemens [1925] of a rare hereditary skin disorder known as
keratosis follicularis spinulosa decalvans in a large family from
Munich. The affected males in this family had extensive
follicular keratosis, alopecia of the eyebrows, eyelashes, beard
and neck, and corneal clouding secondary to eyelid inflamma-
tion. Their female relatives were considered normal, but on
examining them he found more limited (and variable) follicular
keratosis with no alopecia or abnormalities of the eyes or
eyelids.
Siemens at first dismissed the possibility of sex-linked domi-
nant inheritance ‘‘since the existence of this mode of inheri-
tance in human diseases seemed doubtful to me.’’ He adds that
the few disorders put forward at the time as possible examples,
such as ‘‘manic depressive madness,’’ were suggested based
only on their greater frequency in females. Yet his further
studies of this large family showed that all five daughters of
affected males were affected, the seven sons were all free of
disease, and the sons of affected women could be either affected
or normal, leading him to conclude that this was indeed an
example of sex-linked dominant inheritance. But the hair and
eyelid affections were missing in females, leading him to con-
clude ‘‘it is incomplete dominant sex-linked inheritance.’’
Siemens goes on to underscore the entire problem with
X-linked inheritance that we are addressing again 75 years
later by commenting that ‘‘One could freely reflect and
approach this phenomenon also from the other side and say
the disorder is recessive indeed, but the recessiveness is,

contrary to the other sex-linked inherited disorders, not
complete since heterozygote women show signs of their genetic
affection’’ [Siemens, 1925]. Recognizing the difficulty, Siemens
proposes criteria to differentiate incomplete forms of sex-
linked recessive and dominant inheritance: ‘‘For the label of
incomplete dominance one must demand (1) that when possible
all heterozygotes are (incompletely) affected and (2), that the
affection of the heterozygotes is easily ascertainable, also
clinically apparent’’ [Siemens, 1925]. However, we find these
criteria difficult to apply in practice. For example, completely
unaffected heterozygotes occur in most so-called X-linked
dominant disorders.
Evolving Concepts of X-linked
Inheritance in Humans
The conceptual basis of heredity elucidated by Morgan and
others soon appeared in standard texts on genetics, where the
terminology changed from sex linked to the current X-linked
inheritance. By 1931, Baur et al. [1931] postulated that sex-
linked dominance must exist about twice as often in females as
in males and that recessive hereditary factors could only mani-
fest themselves in males. Jennings extended this by writing
that ‘‘dominant effects manifest in all individuals in which the
defective X is present’’ and that ‘‘recessive characteristics are
manifested only if a normal X is not present’’ [Jennings, 1935].
Several decades later, human genetics texts narrowed the
definition further so that recessive disorders would only mani-
fest in females with two copies of the mutant gene and in males
if they had one copy [Kemp, 1951; Emery, 1968]. Considered
together, these brief extracts illuminate a gradual change in
use of the terms dominant and recessive. Their use has become
far more rigid than in earlier works, as they are being applied to
traits without regard to the specific experimental situation or
family under observation.
Even recent texts on human genetics state that males are
affected almost exclusively (or ‘‘mainly’’) with X-linked reces-
sive disorders, which are transmitted through unaffected
carrier females to their sons (Table I), save for a few females
who may be affected based on the pattern of X inactivation.
Most also state that daughters of affected males with X-linked
dominant disorders always inherit the disorder, and that
affected females typically have milder, although variable,
expression of the phenotype [Rimoin et al., 1997; Strachan and
Read, 1999; Nussbaum et al., 2001].
METHODS
Both this historical perspective on X-linked inheritance and
our personal experience with human X-linked disorders led
us to question current thinking regarding existing rules for
X-linked diseases. We therefore conducted a review of the
TABLE I. Summary of Rules for X-Linked Inheritance
Summarized From Rimoin et al. [1997] (pp 93–94)
X-linked recessive
Males are affected almost exclusively
Transmission occurs through unaffected or carrier females to
their sons
Male-to-male transmission is not observed
Affected males are at risk of transmitting the disorder to their
grandsons through their obligate carrier daughters
X-linked dominant
Daughters of affected males always inherit the disorder
Sons of affected males never inherit the disorder
Affected females can transmit the disorder to offspring of both
sexes
An excess of affected females exists in pedigrees for the disorder
X-Linked Inheritance 139
penetrance and expressivity of several X-linked disorders. We
selected more than 40 X-linked disorders for review, preferen-
tially choosing those with a high enough frequency that ob-
servations in large series of patients had been possible, and in
which the gene has been identified and no major autosomal
phenocopy was known. From among these, we were able to
collect data of adequate although variable quality on pene-
trance and expressivity in 32 disorders.
To simplify the analysis, we divided male and female pheno-
types for all disorders into only two groups: severe and mild.
When standard disease classifications used three or more
groups, we combined them as needed until only two groups
were left. For example, males with mutations of the ABCD1
transporter gene may have childhood adrenoleukodystrophy
(ALD), juvenile ALD, typical adrenomyeloneuropathy (AMN),
the cerebral form of AMN, or isolated Addison disease. We
combined childhood and juvenile ALD into a severe group, and
the remaining three subtypes into a mild group. Details re-
garding our classification of all disorders are available on
request.
We then defined penetrance as the proportion of all symp-
tomatic individuals (i.e., those with any expression of disease
but not including those with abnormal test results only) among
carriers of the mutation, both symptomatic and asymptomatic.
We defined a severity index as the proportion of all sympto-
matic individuals with a phenotype classified as severe among
all symptomatic carriers.
RESULTS
Results of our analysis are shown in Table II and Figure 3. As
expected, both penetrance and severity index were much
higher in males than females. In males, penetrance was high
(defined as greater than 90%) in 29 of 32 disorders. The three
disorders with intermediate penetrance (defined as between
11% and 90%) were ALD, glycerol kinase deficiency, and
G6PD-related hemolytic anemia, all thought to be associated
with environmental triggers. Similarly, the severity index was
above 70% in 21 and above 90% in 12 of 32 disorders.
In females, penetrance was high in 9 of 32 disorders or 28%,
intermediate in 10 of 32 disorders or 31%, and low (defined as
0–10%) in 13 of 32 disorders (40%). The 10 disorders with
intermediate penetrance varied widely with a range of 15–85%
and mean of 43.4%, and were without a common characteristic
that we could identify. Most have been classified as X-linked
‘‘recessive’’ and have a low severity index. The severity index
was above 70% in only 3 of the 31 disorders (10%), all three of
which have been considered as classical X-linked ‘‘dominant’’
disorders: pyruvate dehydrogenase E-1-a subunit deficiency,
vitamin D resistant rickets, and the Xg blood group. While the
data on penetrance in females is bimodal, no clear gap between
the peaks exists that would allow separation into dominant
and recessive subgroups.
DISCUSSION
Our reconsideration of X-linked inheritance in humans was
prompted by years of experience evaluating patients and
families with X-linked disorders. Time and again, we found
that the rules differentiating X-linked dominant and recessive
inheritance failed to explain why so many female carriers of
X-linked recessive disorders had an abnormal phenotype,
regardless of severity. We have encountered anecdotal expec-
tations that females with X-linked disease phenotypes should
have unfavorable skewing of X inactivation, but found few
proven examples of this phenomenon other than Duchenne
muscular dystrophy [Worton and Brooke, 1995]. Yet muscle
fibers are multinucleated syncytia that might be expected to
follow different rules than the norm. In our own studies, we
140 Dobyns et al.

TABLE II. Penetrance and Severity Index in 32 X-Linked Diseases in Humansa

Male Female

Disease (gene or protein) Penetrance Severity index Penetrance Severity index

Adrenoleukodystrophy (ABCD1) 90 39 20 Low

Alpha-thalassemia/mental retardation (XH2) 100 100 Low Low
Alport syndrome (COL4A5) 100 100 15 Low
Amelogenesis imperfecta (AMELX) 100 100 100 No data
Androgen insensitivity syndrome (AR) 100 70 10 Low
Chondrodysplasia punctata (ARSE) 100 100 Low Low
Chondrodysplasia punctata (CDPX2) 100 100 93 29
Chronic granulomatous disease (CYBB) High 94 10 Low
Duchenne and Becker muscular dystrophy (dystrophin) 96 65 22 Low
Emery-Dreifuss muscular dystrophy (Emerin) 100 58 Lowb Low
Fabry disease (alpha-galactosidase) 100 84 70 4
Fragile X syndrome (FMR1) >95 87 53 47
Glycerol kinase deficiency 56 78 Low Low
Hemolytic anemia (G6PD) 41 10 Low Low
Hemophilia A (F8) or Hemophilia B (F9) 100 68 10 Low
Hunter syndrome (IDS) 100 81 Low Low
Hydrocephalus with aqueductal stenosis, X-linked 100 97 Low Low
spastic paraplegia (L1CAM)
Ichthyosis, steroid sulfatase deficiency (STS) 100 90 24 Low
Kallmann syndrome (KAL1) 100 83 67 Low
Kennedy disease (AR) 100 10 58 Low
Lesch-Nyhan syndrome (HPRT) 100 76 Low Low
Lissencephaly and subcortical band heterotopia (DCX) 100 66 >90% <5
Lowe oculocerebrorenal syndrome (OCRL1) 100 75 95 Low
Myotubular myopathy (MTM1) 100 95 Low Low
Norrie disease (NDP) 100 30 Low Low
Ocular albinism, Nettleship-Falls (OA1) 100 96 85 Low
Ornithine transcarbamylase deficiency (OTC) 100 50 20 Low
Oro-facial-digital syndrome 1 (OFD1) 100 100 100 33
Periventricular nodular heterotopia (FLN1) 100 <100 95 29
Pyruvate dehydrogenase deficiency (E1a) 100 78 100 72
Vitamin D resistant rickets, hypophosphatemia (PHEX) 100 93 100 80
Xg blood group (PBDX) 100 35 100 100
a
Only symptomatic individuals considered affected (i.e., positive test results only were not sufficient).
b
Eighteen percent of female carriers had ECG abnormalities, but no symptoms.

found a high frequency of post-zygotic mosaicism in heterozyg-

ous mothers of children with mutations of either of the known
X-linked lissencephaly genes, DCX [Gleeson et al., 2000;
Matsumoto et al., 2001] and ARX (W.B. Dobyns, unpublished
data).
Our review of many well-known X-linked disorders showed a
continuum in penetrance between the two sexes. At one end,
female penetrance is low so that nearly all affected individuals
are male. In the middle, penetrance is high in males and
females, although severity is not equal. This produces a female
to male ratio of 2:1, which reflects the ratio of X chromosomes in
the population and thus the opportunity to harbor a mutation
in females and males, respectively. At the other end of the
continuum, penetrance remains high in both sexes but males
die early in gestation, so that nearly all living affected in-
dividuals are female.
We conclude that the rules for X-linked inheritance should
be modified to incorporate our current knowledge of molecular
genetics, and the original intent of the terms insofar as it
is possible to do so. Furthermore, few hypotheses have been
generated to explain differences in the X inactivation pattern
and frequency of mosaicism among different X-linked disorders.

Modified Rules For X-Linked Inheritance

In Humans
Fig. 3. Bar graph shows penetrance of the phenotype in females with
mutations in X-linked genes. The distribution of female penetrance is As usual in science, criticizing current knowledge and
bimodal, although many diseases have intermediate values. practice is easy. The real challenge is to improve them. We
 X-Linked Inheritance 141

have analyzed the existing rules for human X-linked diseases
from many sources and divided them into two groups. The first
consists of strict rules related to segregation of the X and Y
chromosomes in meiosis (Table IIIA). The basis for these rules
has moved beyond proof to scientific dogma.
The second group consists of rules regarding penetrance,
expressivity, and sex ratio that have a complex biological basis,
relating to the hemizygous versus heterozygous state, dosage
compensation by X inactivation, and even parental origin
of spontaneous mutations. In this study, we propose rules
regarding the sex ratio in relatively simple form related only to
penetrance and expressivity (Table IIIB). However, further
revisions to these rules will be needed as our understanding of
X chromosome biology advances.
For example, the parent-of-origin of spontaneous muta-
tions may have an important effect, particularly when female
penetrance is high and reproductive fitness low. The latter
situation also predicts a high rate of spontaneous mutations.
In brief, when a preponderance of spontaneous mutations
occurs in the male germ line, the next generation will have
a predominance of heterozygous females and few hemizyg-
ous males. In contrast, when a predominance of spon-
taneous mutations occurs in the female germ line, the sex
ratio will be about 1:1 if males survive. If mutations cause
prenatal lethality in males, only affected females will be
observed.
The parent-of-origin effect is less significant when female
penetrance is low and reproductive fitness high. When a pre-
ponderance of mutations occurs in the male germ line, the next
generation will have asymptomatic carrier females, while
the following generation will have a predominance of affected
males. When a preponderance of mutations occurs in the
female germ line, all subsequent generations will have a pre-
dominance of affected males.

TABLE III. Revised Rules for X-Linked Inheritance in Humans (Mammals)

A: Rules related to segregation of the X and Y chromosomes
. Hemizygous males transmit X chromosomes to daughters and Y chromosomes to sons
Male-to-male transmission of X-linked disorders cannot occur
Sons of hemizygous males never inherit the disorder
Daughters of affected males are all heterozygous (carriers or affected)
All affected males in a family are related through heterozygous females
. Heterozygous females transmit X chromosomes to both sons and daughters
Fifty percent of sons of heterozygous females will be hemizygous males
Fifty percent of daughters of heterozygous females will also be heterozygous females
B: Rules related to penetrance, expressivity and sex ratio
. Penetrance and expressivity
In general, penetrance is higher and expressivity more severe in hemizygous males than in heterozygous females
. Sex ratio of affected individuals (also see text)
Males are predominately affected when female penetrance is low
Female to male ratio is close to 2:1 when female penetrance and male survival are high
Females are predominately affected with prenatal lethality in males
C: Hypotheses related to other biological mechanisms including cell autonomy or non-autonomy of the gene product, X inactivation
status and mosaicisma
. Group 1: Gene products that are cell autonomous and cause early cell death
Cells with the normal gene inactivated die or fail to divide, so that cells with the normal gene active contribute many more cells to
the developing organism (cellular selection)
Heterozygous females will have favorable skewing of Xi based on cell selection; some will have a normal phenotype while others
may become less severe over time
Possible example: incontinentia pigmenti
. Group 2: Gene products that are cell autonomous but do not cause early cell death
Cells with the normal gene active and cells with the mutant gene active both contribute to the developing organism
Fifty percent of cells will have less or no functional gene product and 50% will have normal gene product; the abnormal cells
frequently result in penetrance of the phenotype
Heterozygous females ascertained because of an abnormal phenotype will have germline mutations and random Xi patterns
Heterozygous females ascertained via relatives other than their children will usually have germline mutations and random Xi
patterns
Heterozygous females ascertained as mothers of affected children will frequently have favorable skewing of Xi or post-zygotic
mosaicism
Possible examples: Most X-linked diseases should fall into this category
. Group 3: Gene products that are cell non-autonomous or functionally non-autonomous
Cells with the mutant and normal genes active both contribute to the developing organism
All cells will have, on average, 50% of the normal gene product or the functional effects of 50% of the normal gene product
Heterozygous females ascertained by family history will usually have a normal phenotype, random Xi, and germline mutations
Heterozygous females ascertained because of an abnormal phenotype will have unfavorable skewing of Xi
Non-autonomous gene products would include primarily secreted proteins, while functionally non-autonomous gene products
would include proteins expressed in non-clonal, multinucleated cells such as muscle fibers, and proteins that catabolize small
molecules that pass freely between cells
Possible examples: Duchenne muscular dystrophy (expressed in multinucleated cells), hemophilia (secreted protein),
ornithine transcarbamylase deficiency (acts on substrate that passes freely from cell to cell)
. Group 4: Gene products that directly affect X inactivation (directed skewing)
Heterozygous females will frequently have skewed Xi, with the direction of skewing related to the mutational mechanism
Heterozygous females with an abnormal phenotype will usually have unfavorable skewing of Xi
Heterozygous females with a normal phenotype will usually have favorable skewing of Xi
Example: Skewing of Xi due to mutations of XIST
a
As a reminder, these are given as hypotheses in need of scientific validation, not rules. Partial escape from X inactivation (leaky Xi) in females should have
the same affect as somatic mosaicism in ameliorating the phenotype.
142 Dobyns et al.
While these rules are more widely applicable than existing
rules that differentiate dominant and recessive subtypes,
we think they are still incomplete, as they fail to explain some
aspects of X-linked inheritance. The remaining questions are
clear. Why are heterozygous females sometimes affected and
sometimes not? What factors are responsible for the variable
expression seen in female heterozygotes for X-linked disor-
ders? Using labels such as dominant and recessive answers
neither of these questions, nor do any existing rules or our
modified rules for X-linked inheritance.
To address these questions, we have developed a preliminary
set of hypotheses that involve the biological mechanisms of cell
autonomous versus non-autonomous gene expression, random
versus skewed X inactivation (Xi), and germline versus mosaic
occurrence of mutations (Table IIIC). For this purpose, we will
define gene products as cell autonomous if they function
independently of other cells within the tissue or organism, and
as non-cell autonomous if they function in concert with gene
products from other cells or nuclei. Most of these hypotheses
and the Group 2 modified rules should be amenable to ex-
perimental confirmation.
Proposed New Rules Based on
X Chromosome Biology
In all mammals including humans, the dosage of genes on
the X chromosome is balanced or ‘‘compensated’’ between
males and females by inactivating one of the two X chromo-
somes in every female cell, a process known as X chromosome
inactivation or Lyonization [Lyon, 1961, 1962]. As a result,
normal females are mosaic for X-linked gene expression, with
one population of cells expressing genes from the maternal X
chromosome and the other expressing genes from the paternal
X chromosome. The relative ratio of these two cell populations
in a given female is known as the Xi pattern. Most normal
females have a ratio close to 50:50. An abnormal Xi pattern has
most often been defined as greater than 80:20, although a
stricter value of 90:10 has also been used [Nance, 1964;
Willard, 2000]. For X-linked diseases, favorable skewing
occurs when the normal allele is preferentially active, while
unfavorable skewing occurs when the mutant allele is
preferentially active.
In some older studies and texts, rare affected females with X-
linked ‘‘recessive’’ disorders were explained by either an
X;autosomal translocation or by skewing of Xi, presumably
unfavorable skewing. Both mechanisms have been demon-
strated for Duchenne muscular dystrophy [Worton and
Brooke, 1995]. However, more recent studies have shown
favorable skewing of X inactivation in many other X-linked
conditions including X-linked mental retardation syndromes
[Belmot, 1995; Puck and Willard, 1998; Willard, 2000; Plenge
et al., 2002]. The favorable skewing has been attributed to
secondary cell selection following initially random X inactiva-
tion. Cell selection was proposed to occur against those cells in
which the mutant allele was preferentially active.
Thus, both favorable and unfavorable skewing of Xi have
been found in heterozygous females with X-linked mutations.
These differences are likely to be important, but have remained
largely unexplored. To complicate matters further, one recent
study on this topic contains a conceptual error [Plenge et al.,
2002]. This study reported Xi studies in 24 families with differ-
ent types of X-linked mental retardation, who were selected
‘‘solely on the basis of the clinical presentation in affected males
and therefore represent an apparently unbiased collection of
carrier female subjects.’’ We disagree. The authors assumed
an unbiased sample even though a significant proportion of
female subjects were mothers of affected boys. They have
therefore been selected for the ability to conceive, bear, and
presumably raise children. As a result, these mothers cannot
be treated as an unbiased sample in this context. Females with
phenotypes associated with reduced reproductive fitness (and
fewer or no offspring) are less likely to be ascertained, unless as
sisters, maternal aunts, or other female relatives. This bias
undermines their conclusion that many X-linked mental re-
tardation syndromes are associated with early cell selection.
The same result of frequent skewing in this cohort would result
from ascertainment bias for high-functioning (normal) females
in disorders in which random Xi is associated with frequent
disease expression and reduced reproductive fitness. These
two alternative explanations may be distinguished based on Xi
studies in other female relatives who were not ascertained as
mothers.
From our review of X chromosome biology, we conclude that
randomly unfavorable skewing and cell selection resulting in
favorable skewing are insufficient to explain skewing of Xi in
most circumstances, or the variable penetrance found in so
many X-linked conditions. Any future understanding of female
penetrance must take into account other factors, such as
whether or not the gene product functions in a cell autonomous
manner, whether and when cell selection occurs as the effects
could be very different at different times and locations, and
whether the gene product is directly involved in the process of
Xi or not. Relying on our best understanding of X chromosome
biology, we propose a set of testable hypotheses that may
lead us toward a final set of rules for X-linked inheritance
(Table IIIC).
In closing, we urge that the descriptive terms ‘‘dominant’’
and ‘‘recessive’’ be dropped from use with any X-linked
disorder, that future clinical reports of X-linked disorders take
care to describe the female phenotype clearly, and that existing
and possible new rules for X-linked inheritance be subjected to
experimental study.
ACKNOWLEDGMENTS
We thank the many families with X-linked disorders whom
we have met in clinics and have participated in research
projects over the years, as their experiences have taught us
most of what we know of these disorders, and the entire BioSci
21200 class in 2002 who began this study. We also thank
Carolyn Schanen for review of the paper and helpful comments
regarding X-linked inheritance, and Jennifer Moran and
Jonathan Rotter for assistance with the Talmudic citation.
REFERENCES
Babylonian Talmud. Yevamot 64b. Babylonian Talmud.
Baur E, Fischer E, Lenz F. 1931. Human heredity. In: Paul E, Paul C,
editors. New York: Macmillan Company.
Belmot JW. 1995. Insights into lymphocyte development from X-linked
immune deficiencies. Trends Genet 11:112–116.
Benson KR. 2001. Morgan’s resistance to the chromosome theory. Nat
Reviews Genet 2:469–474.
Conklin EG. 1915. Heredity and environment in the development of men.
Princeton: Princeton University Press.
Emery AEH. 1968. Heredity, disease and man: Genetics in medicine.
Los Angeles: University of California Press.
Gleeson JG, Minnerath S, Kuzniecky RI, Dobyns WB, Young ID, Ross ME,
Walsh CA. 2000. Somatic and germline mosaic mutations in the
doublecortin gene are associated with variable phenotypes. Am J Hum
Genet 67(3):574–581.
Horner JF. 1876. Die Erblichkeit des Daltonismus. Amtlicher Bericht ueber
die Verwaltung des Medizinalwesens des Kantons Zurich vom Jahr 1876
Zurich: Druck der Genossenschaftsbuchdruckerei. pp 208–211.
Jennings HS. 1935. Genetics. New York: W.W. Norton & Company.
Kelley RL, Wang J, Bell L, Kuroda MI. 1997. Sex lethal controls dosage
compensation in Drosophila by a non-splicing mechanism. Nature
387(6629):195–199.

Kemp T. 1951. Genetics and Disease. Copenhagen: Ejnar Munksgaard.
Keyes LN, Cline TW, Schedl P. 1992. The primary sex determination signal
of Drosophila acts at the level of transcription. Cell 68(5):933–943.
Lyon MF. 1961. Gene action in the X-chromosome of the mouse
(Mus musculus L.). Nature 190:372–373.
Lyon MF. 1962. Sex chromatin and gene action in the mammalian X
chromosome. Am J Hum Genet 14:135–148.
Matsumoto N, Leventer RJ, Kuc JA, Mewborn SK, Dudlicek LL, Ramocki
MB, Pilz DT, Mills PL, Das S, Ross ME, et al. 2001. Mutation analysis of
the DCX gene and genotype/phenotype correlation in subcortical band
heterotopia. Eur J Hum Genet 9(1):5–12.
Morgan TH, Sturtevant AH, Muller HJ, Bridges CB. 1915. The mechanism
of Mendelian heredity. New York: Henry Holt & Company.
Morgan TH, Sturtevant AH, Muller HJ, Bridges CB. 1922. The mechanism
of Mendelian heredity. New York: Henry Holt & Company.
Nance WE. 1964. Genetic tests with a sex-linked marker: Glucose-6-
phosphate dehydrogenase. Cold Spring Harbor Symp Quant Biol 29:
415–424.
Nussbaum RL, McInnes RR, Willard HF, Boerkoel CFr. 2001. Thompson &
Thompson: Genetics in medicine. Philadelphia: W.B. Saunders Co.
Parkhurst SM, Meneely PM. 1994. Sex determination and dosage com-
pensation: Lessons from flies and worms. Science 264(5161):924–932.
Parkhurst SM, Bopp D, Ish-Horowicz D. 1990. X:A ratio, the primary sex-
determining signal in Drosophila, is transduced by helix-loop-helix
proteins. Cell 63(6):1179–1191.
X-Linked Inheritance 143
Plenge RM, Stevenson RA, Lubs HA, Schwartz CE, Willard HF. 2002.
Skewed X-chromosome inactivation mental retardation disorders. Am J
Hum Genet 71:168–173.
Puck JM, Willard HF. 1998. X inactivation in females with X-linked disease.
N Engl J Med 338(5):325–328.
Rimoin DL, Connor JM, Pyeritz RE. 1997. Emery and Rimoin’s Principles
and Practice of Medical Genetics. New York: Churchill Livingstone.
Siemens HW. 1925. Ueber einen, in der menschlichen Pathologie noch nicht
beobachteten Vererbungsmodus: Dominant-geschlechtsgebundene Ver-
erbung. Archiv fur Rassen und Gesellschafts-Biologie; einschliesslich
Rassen und Gesellschafts hygiene 17:47–60.
Sinclair AH, Berta P, Palmer MS, Hawkins JR, Griffiths BL, Smith MJ,
Foster JW, Frischauf AM, Lovell-Badge R, Goodfellow PN. 1990. A gene
from the human sex-determining region encodes a protein with homol-
ogy to a conserved DNA-binding motif. Nature 346(6281):240–244.
Strachan T, Read AP. 1999. Human molecular genetics 2. New York:
John Wiley & Sons, Inc.
Thompson HS. 1986. Johann Friedrich Horner (1831–1886). Am J Ophthal
102:792–795.
Willard HF. 2000. The sex chromosomes and X chromosome inactivation. In:
Scriver CR, Beaudet AL, Sly WS, Valle D, Childs B, Vogelstein B, editors.
The metabolic and molecular bases of inherited Disease. 8th edition.
New York: McGraw-Hill. pp 1191–1221.
Worton RG, Brooke MH. 1995. The X-linked muscular dystrophies
(Chapter 140). In: Scriver CR, Beaudet AL, Sly WS, Valle D, editors.
New York: McGraw-Hill, Inc. pp 4195–4226.