On Lumpers and Splitters, or the Nosology of Genetic Disease*

Vicror A. McKusick, M.D.**

 

Introduction

Nosology, study of the classification of disease, is
science in terms of the definition of science attributed to
Einstein: “an attempt to make the chaotic diversity of
our sense experience correspond to a logically uniform
system of thought.”

As Knut Faber pointed out in his Nosography, men-
delism contributed much to nosology by focusing atten-
tion on specific entities. Bacteriology, with its similar
emphasis on specific etiology and specific entities, had
a comparable effect. One need only recall that a century
ago, symptoms such as jaundice, dropsy and anemia
were viewed as entities, in much of medicine at least, to
realize the nosologic contributions of mendelism and
bacteriology. But the main object here is to review some
contemporary problems in the nosology of genetic dis-
ease.

The two leading principles in genetic nosology are
pleiotropism and genetic heterogeneity. Pleiotropism
means multiple effects of a single etiologic factor, eg a
single gene in the genetic use to which the term is
usually put. Pleiotropism is the usual, but not sole, basis
for hereditary syndromes. Linkage, that is, close situa-
tion on the same chromosome of genes for the separate
manifestations, is a theoretically unsatisfactory and as
yet unproved explanation for mendelizing syndromes.
Because of crossing-over, which in time separates even
closely situated loci, linkage produces no permanent
association of traits in a population.

Genetic heterogeneity refers to the existence of two or
more fundamentally distinct entities with essentially
one and the same phenotype. Nosologists in all fields

*An article on this same subiect by Doctor McKusick
will appear in Perspectives in Biology and Medicine, Vol.
12, No. 2, Winter, 1969.

**Professor of Medicine, The Johns Hopkins University
School of Medicine, Baltimore.

tend to be either ‘“lumpers” or “splitters” (Fig. 1 a and
b). Psychologists tell us that we find it easier to recog-
nize similarities than differences. Hence a natural ten-
dency to lumping exists. However, geneticists are forced
to be splitters because of their recurrent encounters with
genetic heterogeneity in recent years.

Pleiotropism is “many from one” — multiple pheno-
typic features from one etiologic factor, one gene. Ge-
netic heterogeneity is ‘one from many” — one and the
same or almost the same phenotype from several differ-
ent etiologic factors.

In medical genetics awareness of plelotropism and
genetic heterogeneity have developed, particularly in
the last 20 years, the second a bit later than the first.
Looking back on my own work in medical genetics, I
recognize that pleiotropism was a leading concern in its
earlier stages when I was working, for example, with the
Marfan syndrome. Latterly, genetic heterogeneity has
become increasingly the focus, eg in studies of the
genetic mucopolysaccharidoses and the separation of
homocystinuria from the Marfan syndrome. Successive
editions of Heritable Disorders of Connective Tissue?
illustrate this trend in nosology.

In an earlier period medical genetics suffered from
excessive and improper splitting, which was at least
partly inadvertent, arising as it did from specialization
in medicine. Seeing cases of one and the same entity,
physicians in different specialties were concerned main-
ly with features falling within their particular purview
and often failed to recognize that the feature of partic-
ular interest to them was merely part of a syndrome.
Examples are angioid streaks of the ophthalmologist and
pseudoxanthoma elasticum of the dermatologist. Wer-
mer’ described the syndrome of familial endocrine ade-
nomatosis in 1954. Zollinger and Ellison*? described
“their” syndrome in 1955. Recognition that they are one
has subsequently developed.5° Thus, medical geneticists
have been, and continue to be, lumpers to the extent that

Birth Defects: Original Article Series. Vol. V, No. 1; January, 1969 23
ON LUMPERS AND SPLITTERS, OR THE NOSOLOGY OF GENETIC DISEASE

 

 

Fig. la and 1b. The resemblance of the splitter to the revered Civil War President is a clue

i

to the author’s bias. The cartoons also indicate the splitting is harder work than lumping.

they pull together the pleiotropic manifestations of ge-
netic syndromes. (Indeed, medical geneticists can be
the generalists of modern medicine.) As it was put by
Jonathan Hutchinson (1828-1913), a pioneer syndrom-
ologist, ‘We must analyze, and seek to interpret part-
nerships in disease.’’34

Classification of disease has both similarities to and
differences from taxonomy of plants and animals.53 Tax-
onomists like nosologists tend to be either lumpers or
splitters. But the principal, almost the only, question the
nosologist asks is whether syndromes A and B are one
and the same entity or separate ones. The taxonomist,
on the other hand, constructs a branching classification
based on his interpretation of phylogeny. The compon-
ents in his classification bear varying degrees of genetic
relationship to each other, based on descent from a com-
mon primitive ancestor.54 Although classifications of dis-
eases in particular categories, eg classification of hand-
anomalies, are compiled and have mnemonic and heur-
istic uses, no more than an artificial relationship exists
between the several components.

The Genetic Entity

What constitutes an entity? How does one identify
genetic heterogeneity? Few would quibble with the
statement that the phenotype resulting primarily from
a specific and unitary factor is an entity. Thus, delinea-
tion of genetic entities is on surest ground if a funda-
mental biochemical defect or chromosomal fault is iden-
tified. Trisomy D, trisomy E and cri-du-chat syndromes

24

are now recognizable with a high order of accuracy on
clinical grounds alone. Demonstration of a specific etio-
pathogenetic basis permits separation of groups of
cases in which the frequency of clinical features can be
determined and with which new cases can be compared,
like an “unknown” with a “standard.” One can then
make use of Bayesian methods by which individual
manifestations are given weightings according to their
frequency in the specific entity on the one hand, and in
persons without the specific entity on the other hand. Of
course, the Down syndrome was diagnosable with con-
siderable accuracy before the era of chromosomology
and the Walker scoring method for use in connection
with dermatoglyphic diagnosis was already developed,
but the specificity of the diagnostic technic was im-
proved by working from karyotypically proved cases.
Greater difficulty is encountered in delineating entities
among congenital disorders when the etiology is not
known or a unitary biochemical defect has not been
shown and, consequently, unitary etiopathogenesis is
not certain. Such is the problem, for example, with the
Cornelia de Lange syndrome, the Turner phenotype
with normal karyotype, the Rubinstein syndrome,
Prader-Willi syndrome, etc. In these, overdiagnosis is a
risk. Unless the clinical features are unusually distinc-
tive, criteria are likely to be either too rigorous or too
inclusive and a happy middle grotind is not easily at-
tained. When one does not have a specifically diagnostic
laboratory test or other diagnostic means independent of
the clinical phenotype, one cannot list the frequency of
occurrence of specific features without falling into circu-
Victor A. McKusick, M.D.

 

lar reasoning. (Small chromosomal aberrations beyond
the limits of resolution of presently existing cytogenetic
methods seem likely as the basis of many syndromes
which show rather remarkable case-to-case reproduc-
ibility but which do not show sufficiently strong familial
aggregation to be considered mendelian and do occa-
sionally show detectable chromosomal changes. But un-
til, by improved methods, small aberrations are actually
demonstrated, such is only speculation.) At the stage
when homocystinuria had not yet been separated from
the Marfan syndrome, what was the usefulness of tab-
ulating precise percentage figures for the frequency of
various clinical features? Even today the lack of a spe-
cific diagnostic test in the Marfan syndrome impedes
delineation at the mild end of the spectrum of severity
and the possibility of residual heterogeneity limits the
value of quantitative statements on the percentage fre-
quency of components.

How Does One Identify Genetic Heterogeneity?

The methods for recognizing genetic heterogeneity,
the existence of separate entities in what has been con-
sidered a single phenotype, are mainly three: clinical,
genetic and biochemical.

Fig. 2. The principle of genetic heterogeneity is illustrated by a consideration of this photo-

Clinical Methods

Differences in the phenotype are the most treacherous
basis for decision on genetic heterogeneity. After all,
similarity of phenotype is what leads to mistaken im-
pressions of homogeneity (or unity) in the first place.
Furthermore, genetic disorders, particularly those in-
herited as autosomal dominants, vary widely in severity,
presumably in large part because of effects of modifier
genes. Some aspects of a syndrome may be missing in
the individual case. For example, among sibs who in-
herited the Marfan gene from a parent, one may have all
the cardinal features of the syndrome but another may
not show ectopia lentis. ‘On the average, 50% of the
genes of sibs are identical by descent. Differences in the
other half of the genome account for differences in ex-
pression of the single major gene just as placing a plei-
otropic gene on. different genetic backgrounds in mice
can result in suppression of some aspects of the syn-
drome produced by that gene.3

Clinical differences in phenotype can come to the sup-
port of other methods for recognizing genetic hetero-
geneity. For example, when pedigree pattern suggested
the existence of an X-linked form of mucoplyosacchari-

 

graph of the national annual meeting of an organization of dwarfs and midgets. Clearly
if one were to undertake a study of the genetics of short stature, or of the physiologic
defect underlying short stature, or of other aspects including even the psychologic and
sociologic impact of short stature, separation of separate categories would be necessary.
In this absurdly obvious example, genetic heterogeneity is evident even to the casual
observer and the experienced eye can pick out more than a dozen distinct entities. (Of
course, heterogeneity can also involve nongenetic bases for the qhenotype, so called

phenocopies.)

25
ON LUMPERS AND SPLITTERS, OR THE NOSOLOGY OF GENETIC DISEASE

 

dosis (Hunter syndrome), analysis of phenotype in
comparison with the autosomal recessive type showed
several phenotypic differences, particularly absence of
corneal clouding.”

Analysis of phenotype can be most reliably used to
delineate an entity when an extensive kindred with
many persons affected with an autosomal dominant are
available for study, or a deme* (really an extended
kindred) with many cases of an autosomal recessive. If
the disorder is rare ( as are most mendelizing dis-
orders), then one can be confident that the identical
gene is responsible for all cases in a given kindred or
in a given deme. The range of variability in the pheno-
type of one entity can be more securely determined than
is possible on the basis of a large number of randomly
ascertained families each with only a few affected mem-
bers. In the latter situation circularity enters because
of the criteria by which families were selected in the
first place.

Let me illustrate with examples of dominants. The
view that familial multiple polyposis of the colon and
the Gardner syndrome are separate entities has been
challenged by surgeons who, focusing mainly on indi-
vidual cases, feel that they represent a spectrum with
no cutoff point between two distinct entities. The
strongest support for distinctness comes from large
family studies. The most extensively involved family of
multiple colonic polyposis yet studied is, to my knowl-
edge, that of Asman and Pierce with about 90 affected
persons, not one of whom had extra-bowel tumors of
the type seen in the Gardner syndrome.§ The most ex-
tensively involved family reported** with the Gardner
syndrome is, to my knowledge, the one originally studied
by Eldon Gardner. More than 20 persons in this kindred
have the entity although a few seemed at the time of first
study to have only polyps and others seemed to have
only extraintestinal tumors.7

Another dramatic example of the usefulness of study-
ing large families is provided by amyloid neuropathy.
While a research fellow with me from 1965 to 1967,
Mahloudji studied a form of amyloid neuropathy which
was initially ascertained in 11 seemingly unrelated
kindreds living mainly in Maryland and in neighboring
states. Genealogic sleuthing showed that all 11 kin-
dred were descended from a couple who immigrated
from Germany in the 18th century. The giant kindred
contained at least 145 persons who were affected, accord-
ing to reliable information, or must have had the gene
because of their position in the genealogy. Personal
study of over 50 affected persons from this single kin-
dred leaves no doubt that the disorder is distinct from
the amyloid neuropathy observed in equally numerous
cases in Portugal by Andrade.? Table I reviews the
phenotypic differences of the two types which are given
either an eponymous or a geographic name; the ailment
in Mahloudji’s enormous kindred appears to be iden-

*Deme is defined by George P. Murdocké as a local endo-
gamous community, or consanguinal kin group.

**Recent restudy of previously reported polyposis family42
shows that the disorder is in fact the Gardner syndrome
and that about 60 members of the kindred are affected.

26

tical to that studied in Indiana by Rukavina and his
colleagues!9 in a kindred of Swiss extraction.

Wide variability in elliptocytosis is indicated by an
extensive study of that disease in Iceland*’? where all
affected persons can plausibly be considered as having
the same gene, ie as being members of a single large
kindred. Although linkage studies indicate the existence
of at least two genetically distinct forms of elliptocy-
tosis,5! the wide variability in this single kindred coun-
sels caution in interpreting a particular phenotype as
characteristic of the form linked to Rh-type as con-
trasted with the nonlinked form.

The range of variability of the ordinarily very rare
autosomal recessive Ellis-van Creveld syndrome’! (“six-
fingered dwarfs”) has been gauged better from analysis
of the 62 cases which have up to now been recognized in
a single inbred group, the Lancaster County (Pa.)
Amish. The fact, for example, that one-third die before
age six months cannot be determined from analysis of
reported cases (which also number about 60) because
of biases of ascertainment. Sometimes a syndrome ap-
parently inherited as a recessive is observed in two or
more sibs. Especially if the parents are related, the
possibility comes up that the affected children fell heir
to two unrelated recessives rather than the disorders
being a pleiotropic syndrome produced by a single mu-
tant gene. (Close linkage of two recessives would in-
crease the stimulation of pleiotropism.) If a large inbred
group has many individuals showing the coincidence of
the two manifestations, single gene etiology is strongly
supported. Cartilage-hair hypoplasia has been deline-
ated by the fact that over 80 Amish persons (by present
count) have the associated hair and skeletal abnor-
mality.!2 More recently its anticipated occurrence in
non-Amish persons has been documented.44:45 Similarly,
for a dominant such as the Peutz-Jeghers syndrome, in
which the coincidence of such different features as
melanin spots and intestinal polyposis is difficult to

TABLE |

DIFFERENTIAL FEATURES OF TWO TYPES
OF HEREDITARY AMYLOID NEUROPATHY

 

 

Characteristic Type | of Type Il of
Andrade Rukavina
(Portuguese type) (Indiana type)

Age of onset 3rd decade 5th decade
Site of onset Feet Hands
G. |. complaints Common Rare
Impotence and

sphincter

disturbance Common Rare
Foot ulcers Common , Rare
Duration of

Iliness 4-12 years 14-40-++ years

 
Victor A. McKusick, M.D.

 

explain on physiologic grounds, the study of an exten-
sively affected kindred lends strong support to its single
gene basis.13 If separate genes were responsible for the
two features they would become separated at some point
in a large kindred (even if closely linked) through the
process of genetic crossing-over. This is, then, essentially
a genetic approach to definition of a syndromal entity—
which brings us to this aspect.

Genetic Methods

Genetic methods for identifying heterogeneity include
mode of inheritance, allelism test and linkage study. A
number of phenotypes are known which in different
families exhibit autosomal dominant, autosomal reces-
sive and X-linked patterns of inheritance. No one could
consider the autosomal and X-linked forms anything
but separate entities since the mode of inheritance in-
dicates that the genes are located in different parts of
the genome. The distinctness of the autosomal dominant
and autosomal recessive forms is likewise defensible in
many instances. I have had an opportunity to study an
X-linked spastic paraplegia!4 and an autosomal domin-
ant form.15 Autosomal recessive spastic paraplegia with
or without other features has been observed by me, my
colleagues, and by others.1617 Retinitis pigmentosa and
peroneal muscular atrophy were examples of triple mode
of inheritance that William Allan cited when he pointed
out the principle that is sometimes called Lenz’ law.18
The autosomal dominant form is usually the mildest
and the autosomal recessive form the most severe, with
the X-linked recessive form occupying an intermediate
position as to severity.

When two individuals homozygous at the same locus
have children, all their children are likewise homozy-
gous at that locus. Deaf-mutes frequently marry and in
some such families all the children are also deaf-mutes;
the parents must have an allelic form of deafness. But in
others, although both parents seem from the evidence in
their family trees to have a recessively inherited form
of deaf-mutism, and although the disorder in the parents
is phenotypically indistinguishable, all children have
normal hearing. Presumably the parents are homozy-
gous at separate loci; they have nonallelic forms of re-
cessive congenital deafness. Chung, Robinson and Mor-
ton concluded that homozygosity at any one of many
loci may result in congenital deafness of phenotypically
indistinguishable type — a large amount of genetic
heterogeneity.19 I find Morton’s conclusion plausible
because usually when one studies an inbred group for
specific recessives one finds deaf-mutism whatever else
may or may not be present. We have observed 14 cases
of deaf-mutism among the Amish of Lancaster County,
Pa. and 38 cases of deaf-mutism among the conservative
Mennonites of the same county.?° In one instance, an
Amish deaf-mute married a Mennonite deaf-mute. All
three of their children hear normally, suggesting non-
allelic recessive deafness in the two religious isolates.
Among the cases of isolated growth hormone deficiency,
we have identified two genetically distinct although
phenotypically indistinguishable types by the fact that
an affected man and woman had two normal children.?4
Heterogeneity in recessive albinism,?? and in recessive
congenital amaurosis, or retinal aplasia,?> has been
demonstrated by the same approach.

 

BIOGRAPHIC DATA

Dr. Victor A. McKusick

is Professor of Medicine

and Chief, Division of Med-

ical Genetics, Department

of Medicine, The Johns

Hopkins University School
of Medicine.

He was born in Maine,
attended Tufts College and
graduated from The Johns
Hopkins Medica! School in
1946.

He is the.author of
Heritable Disorders of Con-
nective Tissue, Human
Genetics, On the X Chro-
mosome of Man and other books.

 

He has also contributed to cardiology, and wrote
a well-known monograph entitled Cardiovascular
Sound in Health and Disease.

 

 

 

Linkage, the demonstration by family studies that two
genes are reasonably close together on the same chromo-
some pair, can demonstrate heterogeneity. The locus for
one form of elliptocytosis is closely situated to the Rh
blood group locus. On the other hand, a form of ellipto-
cytosis as yet phenotypically indistinguishable is deter-
mined by a gene at a locus which shows no linkage with
the Rh lIocus.24 Even if other methods had not revealed
the heterogeneity in X-linked hemophilia (hemophilias
A and B), sufficiently extensive linkage studies might
have uncovered the heterogeneity; the hemophilia A
locus is closely linked to that for glucose-6-phosphate
dehydrogenase and those for color blindness, but the
hemophilia B locus is so far removed that independent
assortment occurs through crossing-over.25

Biochemical Methods

Demonstration of biochemical differences is in man
the most critical and definitive means of establishing
genetic heterogeneity. The biochemical changes ob-
served may be useful even though moderately far re-
moved from primary gene action. For example, the
qualitative pattern of urinary excretion of mucopoly-
saccharides in the mucopolysaccharidoses helps disting-
uish several types,2 eg types III and IV from each other
and from types I and II, although some are disting-
uished with difficulty on this ground. (eg types I and
II). Differentation is surer when a biochemical defect
closer to primary gene action is demonstrable, eg an
enzyme defect. Firm delineation of multiple forms of
glycogen storage disease and of hereditary nonsphero-
cytic hemolytic anemia is now possible on the basis of
demonstrated defects in different enzymes.?6

Even when deficiency in the activity of a specific
enzyme is demonstrable, residual heterogeneity may
exist. This has been shown to be the case with non-

27
ON LUMPERS AND SPLITTERS, OR THE NOSOLOGY OF GENETIC DISEASE

 

spherocytic hemolytic anemia due to glucose-6-phos-
phate dehydrogenase deficiency,”6 and that due to pyru-
vate kinase deficiency.27 It has also been shown in
galactosemia?§ and in phenylketonuria or at least phenyl-
alinemia.2®9 It is suspected or tentatively demon-
strated in homocystinuria and several other conditions.
This heterogeneity can be allelic; mutation may have
occurred at different points in the cistron so that amino
acid substitution occurs at different places in the enzyme
polypeptide, resulting in several forms of mutant en-
zyme each with deficiency of enzyme activity but with
different reaction characteristics and different degrees
of deficiency. Following the precedent of cystathionin-
uria,?9 and suggestive differences in response to pyrid-
oxine in homocystinuria, one might think that some
mutations in enzymes affect mainly the combination or
collaboration of the enzyme with the cofactor. The net
effect of either type of mutation would be deficiency of
enzyme activity.

Thus, three main approaches are available for de-
lineating separate gerietic entities by uncovering genetic
heterogeneity — clinical, genetic and biochemical. The
complementarity of hemophilias A and B — blood from
patients with hemophilia A corrects the coagulation
defect in patients with hemophilia B and vice versa —
is a physiologic demonstration of heterogeneity, which,
I suppose, can be included in the clinical method. Mor-
phologic methods, used for instance in sorting out
genetic disorders of the central nervous system, can
perhaps be viewed also as a special type of clinical
analysis. Cell culture technics are supplementing in a
powerful manner both morphologic and biochemical
approaches to the unravelling of heterogeneity.3!

A fascinating proof of genetic heterogeneity in the
mucopolysaccharidoses is provided by the demonstra-
tion by Fratantoni et al.43 When fibroblasts from a
patient with the Hurler syndrome (mucopolysacchari-
dosis I) are mixed in culture with those from a patient
with the Hunter syndrome (mucopolysaccharidoses IT)
or the Sanfilippo syndrome (mucopolysaccharidoses
IIT), “cross-correction” occurs. The accumulations of
mucopolysaccharide disappear from the cytoplasm. Fi-
broblasts from a normal person cause clearing in the
fibroblasts of any of the three types.

Practical Difficulties in Entity Recognition

The practical problems associated with recognition of
genetic entities are considerable. The entity may have a
certain Gestalt which is unmistakable, but is scarcely de-
finable on the basis of a single case and it may be diffi-
cult for even the most skilled word-artist to convey the
features to others who may have seen identical cases. In
this field a photograph can be worth a thousand words.
(The Williams facies of the hypercalcemia-supravalvar
stenosis-mental retardation syndrome is a good case in
point.4¢) The entities are individually rare. A single
case may not impress; it may take two to arouse suspi-
cion that a discrete entity is involved and three may be
necessary to convince. Of course, if two or more sibs are
identically affected this helps, but with the small fam-
ilies now the rule a majority of families will have only
one affected child (if the condition is an autosomal

28

recessive) and of course in those conditions which are
sublethal dominants all or almost all cases result from
fresh mutation and are sporadic.

Josef Warkany of Cincinnati points out to me that
among the several hundred thousand patients in institu-
tions for the mentally retarded in this country many as
yet undelineated syndromes exist, each with pheno-
typically more or less striking features, but because each
is an isolated case the physician does not know what to
make of it. The problem is a logistic one, how to get
together the two or three or more cases which may exist
in widely separated institutions. Once an entity is de-
scribed on the basis of two or three cases, other cases
tend to be found rather quickly. Many persons had seen
a case of microcephaly, beaked nose and broad thumbs
and great toes, but not known what to make of it until
Rubinstein and Taybi put the syndrome together.35

Inbred Groups and Nosology

The usefulness of inbred groups in defining the pheno-
typic limits of an entity because of the high probability
that the same gene underlies all cases has already been
discussed. Furthermore, delineation of “new”, ie hither-
to unrecognized, genetic entities with recessive inher-
itance is enhanced in inbred groups. At least three
factors account for increased visibility of recessives in
such groups:

1) Inbreeding increases the number of homozygotes.

2) These groups usually have large families.

3) The “groupness” increases conspicuousness of dis-
orders.

Contrary to a prevalent misconception — that inbreed-
ing causes a build-up of “bad” genes in a population,
consanguinity per se does not change the frequency of
genes in a population. It does change the proportion of
genotypes, increasing the number of homozygous indi-
viduals at the expense of heterozygotes.

Intuitively one can appreciate the fact that the greater
the number of children which are produced by two
parents heterozygous for the same recessive gene, the
greater is the chance that two or more will be affected
with a given entity. This is shown mathematically in
Table II. Of families with two heterozygous parents,
ascertainable because at least one affected child has
been born, 86% will have only the one child affected if
there are only two children in the family. Of families
with three children 73% will have only one child af-
fected. Four-child families will in 62% of instances have
only one child affected. Not until six-child families are
reached will more than half the families have more than
one affected child. When we see in a sibship a single
case of an unusual clinical picture which seems to
represent an entity, we cannot be as certain as we can
be if at least two sibs are identically affected.

The “groupness” increases conspicuousness. A deme
is effectively a large kindred so that the considerations
are similar to those just mentioned. But, in addition,
sociologic uniqueness of the group increases visibility.
The occurrence of multiple cases of an unusual clinical
picture in a sociologic distinctive group like the Amish
Victor A. McKusick, M.D.

 

TABLE Il
When sibships affected by an autosomal recessive dis-
order are ascertained through the presence of at least one
affected child,* what proportion of ascertainable sibships
have only one child affected?

 

1 sib families 100%
2 sib families 86%
3 sib families 73%
4 sib families 62%
5 sib families 52%
6 sib families 43%
7 sib families 36%
8 sib families 30%
9 sib families 24%
10 sib families 20%
11 sib families 16%
12 sib families 13%

 

*Ht is further assumed that ascertainment is complete or at least
unbiased so that a random selection of affected sibships is achieved.

is likely to impress the observer even though few or none
of the cases are in sibs. Thalassemia major occurred in
the Mediterranean basin in large numbers of cases but
the definitive description came from a pediatric hema-
tologist in Detroit, Michigan, who could not help but be
impressed with the unusual disorder which occurred
always in children of Mediterranean extraction.32

The Naming of Genetic Entities
Once recognized, entities present problems in naming
— and names are important. A syndrome has “arrived”
if it has a name. Some syndromes have languished for
an appreciable time for lack of a satisfactory name. An
unfortunate consequence of naming can be the mistaken
impression that we understand the condition. With few
congenital disorders does the state of knowledge permit
a designation based on specific etiology or pathogenetic
mechanism, as in the thalidomide syndrome, the rubella
syndrome, and the trisomy 18 syndrome. Eponyms have
their usefulness as mere labels, with no prejudice as to
_ the nature of the basic defect, but place a strain on the
memory. Both physicians’ names (eg most eponyms)
and patients’ names (eg Hartnup’s disease) are used.
I prefer to say the Marfan syndrome (rather than Mar-
fan’s syndrome) because it makes it clear that the
surname is merely a tag. After all, Marfan described the
skeletal features only.

Use of one facet of the syndrome as a name for the
whole has obvious risks since that feature may occur as
an isolated anomaly or may be a part of other syn-
dromes and, on the other hand, it may be missing in
some cases of the particular syndrome. For these reasons
arachnodactyly is a poor designation for the Marfan
syndrome. Of course, if the component manifestation

selected for designating the whole is a striking, unique
and invariable feature of the syndrome, no problem
arises. Examples are focal dermal hypoplasia, incon-
tinentia pigmenti, osteogenesis imperfecta and chondro-
dystrophia calcificans congenita, although even these
are not immune to nosologic, and consequent termino-
logic, problems.

Ideally, the designation should help one remember
the features of a syndrome. Such is the virtue of the
naming system which has been used especially by the
Gorlin school of syndromologists36: oro-facio-digital,
oculo-auriculo-vertebral, oculo-dento-digital, and so on.
It creates no major problem that the trilogic is appro-
priate to more than one entity; one delineated after the
first is simply called “number 2” (as Rimoin and Edger-
ton37 have done for the OFD syndromes) or given an
entirely different name. When their names are reduced
to initials such as OFD, OAV, ODD and FDH (in the
Gorlin system of nomenclature), congenital malforma-
tion syndromes come to sound like governmental agen-
cies — and perhaps often have other similarities!

Beginning with the glycogen storage diseases and
following with the mucopolysaccharidoses, numbering
of disorders in a particular group of entities has become
one way to cope with the nomenclature problem. It has
proved useful to have eponyms to go along with the
numbers. Thus, glycogenoses I through VI carry epo-
nyms von Gierke, Pompe, Forbes,* Andersen, McArdle
and Hers, respectively. Mucopolysaccharidoses I
through VI carry eponyms Hurler, Hunter, Sanfilippo,
Morquio, Scheie and Maroteaux-Lamy, respectively. In
general, it is easier to remember the eponym than the
number. The eponym conjures up a mental picture of
the case(s) the man described.

Geographic designations have occasionally been used,
eg the Portugese and Indiana varieties of amyloidosis
(Table I) and the malformation syndromes called
Amsterdam (Cornelia de Lange syndrome) and Rostock
types.

Although a few simple guidelines for naming of ge-
netic entities can be laid down, usage in the long run
dictates the designation.

The Numerologic Status of Genetic Nosology

The catalogs of mendelian traits?6 which, with com-
puter aids, I maintain on a continuously updated basis
permit an estimate of how many traits are known in
man. The numbers presumably relate to separate loci;
variant hemoglobins with a change in the beta poly-
peptide chain are, for example, listed as one item. Some
allelic traits may be separately listed but these must be
more than counterbalanced by others where nonallelic
heterogeneity has not been recognized. The number of
firmly established traits are shown in Table III, as well
as the number of traits included only provisionally
either because the particular mode of inheritance is not
considered proved or the separateness from other listed
entities is not certain. The well-established loci in man,

*A problem with eponyms is indicafed by the fact that
someS2 call glycogenosis III Cori’s disease.

29
ON LUMPERS AND SPLITTERS, OR THE NOSOLOGY OF GENETIC DISEASE

 

numbering about 700, probably represent only a fraction
of the whole. Man may have 100 times as many genes
in all, perhaps even more than that.41

Probably, proportionately, many more recessives than
dominants remain to be discovered. The predominance
of known dominants over recessives in man is an anoma-
lous situation in comparison with other species, eg the
laboratory mouse, when more recessives than dominants
are known (Table IIIb). The difference is mainly due to
differences in mating patterns. Most visible mutations
(mutations with effects evident without refined methods
for studying the phenotype) are probably recessive. In
man such a mutant gene can arise and be lost—either by
chance or because of some disadvantage even in the
heterozygote—without meeting up with itself in a homo-
zygote. Or if a homozygote does occur, it has a strong
likelihood of being a sporadic case because of the small
size of human families (Table II), and the isolated case
may escape recognition as a distinct entity. On the other
hand, in mice because of brother-sister and other close
matings, as well as the larger number of offspring, a
recessive mutation is likely to come more promptly to
attention.

TABLE fil
NUMBER OF MUTANT PHENOTYPES
ACCORDING TO MODE OF INHERITANCE

 

a. MAN b. MOUSE

 

Verschner (1959)38 McKusick (1968)39 Green (1967)4°

Autosomal

dominant 285 344 (+449) * 99
Autosomal

recessive 89 280 (+349) 207
X-linked 38 68 (+55) 12

 

*The figures in brackets indicate the number of additional traits for
which the particular mode of inheritance has been suggested but not
Proven.

Summary

In genetic nosology both lumping and splitting have a
place: lumping in connection with pleiotropism; split-
ting in connection with genetic heterogeneity.
Victor A. McKusick, M.D.

 

REFERENCES

. Faber, K. H.: Nosography. (2nd ed.), Paul B. Hoeber, Inc.,
1930.

. McKusick, V. A.: Heritable Disorders of Connective Tissue.

(1st, 2nd, ‘3rd ed.), C. V. Mosby Co., 1956, 1960 and 1966.

. Dunn, L. C. and Charles, D. R.: Studies on spotting pat-

terns. J. Analysis of quantitative variations in the pied spot-
ting of the house mouse. Genetics 22:14, 1937.

4. Murdock, G. P.: Social Structure. Macmillan Co., 1948.
5. Smith, W. G. Dis. Colon Rect. Jan.-Feb., 1968.
6. Asman, H. B. and Pierce, E. R.: Louisville, Ky. Unpub-

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

lished observations.

. Gardner, E. J.: Utah State University, Logan, Utah. Per-

sonal communication, 1968.

. Mahloudji, M.; Teasdall, R. D.; Adamkiewicz, J. J.; Hart-

mann, W. H.; "Lambird, Pp. A. and McKusick, V. A:: The
genetic amyloidoses . with particular reference to hereditary
neuropathic amyloidosis, type If (Indiana or Rukavina
type). Medicine, Jan., 1968.

. Andrade, C.: A peculiar form of peripheral neuropathy:

familial atypical generalized amyloidosis with special in-
volvement of peripheral nerves. Brain 75:408-427, 1952.

Rukavina, J. G.; Block, W. D.; Jackson, E. C.; Falls, H. F.;
Carey, J. H. and Curtis, A. C.: Primary systemic amyloid-
osis: a review and an experimental, genetic and clinical study
of 29 cases with particular emphasis on the familial form.
Medicine 35:239-334, 1956.

McKusick, V. A.; Egeland, J. A.; Eldridge, R. and Krusen,
D. E.: Dwarfism in the Amish. I. The Ellis-van Creveld
syndrome, Bull. Johns Hopk. Hosp. 115:306-336, 1964.

McKusick, V. A.; Eldridge, R.; Hostetler, J. A.; Egeland,
J. A. and Ruangwit, U.: Dwarfism in the Amish. II. Cartilage-
hair hypoplasia. Bull. Johns Hopk. Hosp. 116:285-326, 1965.

McAllister, A. J.; Hicken, N. F.; Latimer, R. G. and Con-
don, V. R.: Seventeen patients with Peutz-Jeghers syndrome
in four generations. Amer. J. Surg. 114:839-843, 1967.

Johnston, A. W. and McKusick, V. A.: A sex-linked re-
cessive inheritance of spastic paraplegia. Amer. J. hum.
Genet. 14:83-94, 1962,

McKusick, V. A.: The distribution of certain genes in the
Old Order Amish. Cold Spr. Harb. Symp. quant. Biol.
39:99-114, 1964.

Cross, H. E. and McKusick, V. A.: The Mast syndrome, a
recessively inherited form of presenile dementia with motor
disturbances. Arch. Neurol. 16:1-13, 1967.

Cross, H. E. and McKusick, V. A.: The Troyer syndrome.
A recessive form of spastic parapelgia with distal muscle
wasting. Arch. Neurol. 16:473-485, 1967.

Allan, W.: Relation of hereditary pattern to clinical severity
as illustrated by peroneal atrophy. Arch. intern. Med.
63:1123-1131, 1939.

Chung, C. S.; Robinson, O. W. and Morton, N. E.: A note
of deafmutism. Ann. hum. Genet. 23:357-366, 1959.

Mengel, M. C.; Konigsmark, B. W. and McKusick, V. A.:
Two types of congenital recessive deafness. Eye, Ear, Nose
Thr. Monthly. In press.

Rimoin, D. L.; Merimee, T. J. and McKusick, V. A.: Un-
published observations.

Trevor-Roper, P. D.: Marriage of two complete albinos with
normally pigmented offspring. Brit. J. Ophthal. 36:107-110,
1952 and Proc. roy. Soc. Med. 56:21-24, 1963.

23. Waardenburg, P. J. and Schappert-Kimmijser, J.: Genetics

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

and Ophthalmology, Waardenburg, P. J.; Franceschetti, A.
and Klein, D., Charles C. Thomas, 2:1959 only, 1963.

Morton, N. E.: The detection and estimation of linkage
between the genes for elliptocytosis and the Rh blood type.
Amer. J. hum. Genet. 8:80-96, 1956.

McKusick, V. A.: On the X Chromosome of Man. Ami.
Inst. Biol. Sci., 1964.

McKusick, V. A.: Mendelian Inheritance in Man. Catalogs
of Autosomal Dominant, Autosomal Recessive and X-
linked Phenotypes. (2nd ed), Johns Hopkins Press, 1968.

Bigley, R. H. and Koler, R. D.: Liver pyruvate kinase
(PK) isozymes in a PK- deficient patient. Ann. hum. Genet.
31:383-388, 1968.

Hsia, D. Y.-Y.: Clinical varients of galactosemia. Me-
tabolism 16:419-437, 1967. .

Auerbach, V. H.; DiGeorge, A. M. and Carpenter, G. G.:
Phenylalinemia. A study.of the diversity of disorders which
produce elevation of blood concentrations of phenylalanine.
Amino Acid Metabolism and Genetic Variation. Nyhan,
W. L. Ed., McGraw-Hill, pp. 11-68, 1967.

Frimpter, G. W.: Cystathioninuria: nature of the defect.
Science 149:1095-1096, 1965.

Danes, B. S. and Bearn, A. G.: Cellular metachromasis,
a genetic marker for studying the mucopolysaccharidoses.
Lancet i:241-243, 1967.

Cooley, T. B. and Lee, D.: Series of cases of splenomegaly
in children with anemia and peculiar bone changes. Trans.
Amer. pediat. Soc. 37:29, 1928.

Cooley, T. B.: Erythroblastic anemia. Amer. J. Dis. Child.
43:705, 1932.

McKusick, V. A.: The clinical observations of Jonathan
Hutchinson. Amer. J. Syph. 36:101-126, 1952.

Rubinstein, J. H. and Taybi, H.: Broad thumbs and toes
and facial abnormalities. Amer. J. Dis. Child. 105:588-608,
1963.

Gorlin, R. J. and Pindborg, J. J.: Syndromes of the Head
and Neck, Blakiston Division, McGraw-Hill, 1964.

Rimoin, D. L. and Edgerton, M. T.: Genetic and clinical
heterogeneity in the oral-facial-digital syndromes. J. Pediat.
71:94-102, 1967.

Verschner, O. Fr.v.: Genetik des Menschen. Lehrbuch der
Humangenetik. Urban & Schwarzenberg, Munchen, Berlin,
1959.

McKusick, V. A.: Mendelian Inheritance in Man. Catalogs
of Autosomal Dominant, Autosomal Recessive and X-linked
Phenotypes. 2nd Ed. Johns Hopkins Press, 1968.

Green, M.: The Jackson Laboratory, Bar Harbor, Me.:
Personal communication, August, 1967.

Vogel, F.: A preliminary estimate of the number of human
genes. Nature 201:847, 1964.

Kelly, P. B. and McKinnon, D. A.: Familial multiple poly-
posis of the colon: review and description of a large
kindred. McGill med. J. 30:67-85, 1961.

Fratantoni, J. C.; Hall, C. W. and Neufeld, E. F.: The
defect in Hurler’s and Hunter’s syndromes: family degrada-
tion of mucopolysaccaride. Proc. nat. Acad. Sci. 60:699-
706, 1968.

Beals, R. K.: Cartilage-hair hypoplasia. A case report. J.
Bone Jt. Surg. 50-A:1245-1249, 1968.

31
ON LUMPERS AND SPLITTERS,

OR THE NOSOLOGY OF GENETIC DISEASE

 

45,

46.

47.

48.

49.

50.

Lenz, W.: Diagnosis in Medical Genetics. Crow, J. F. and
Neel, J. V., eds. Proc. Third Int. Congress of Human
Genetics. Johns Hopkins Press. Pp. 29-36, 1967.

Williams, J. C. P.; Barratt Boyes, B. G. and Lowe, J. B.:
Supravalvular aortic stenosis Circulation 2431311, 1961.

Jensson, O.; Jonasson, T. and Olafsson, O.: Hereditary el-
liptocytosis in Iceland. Brit. J. Haemat. 13:844-854, 1967.

Wermer, P.: Genetic aspects of adenomatosis of endocrine
glands. Amer. J. Med. 16:363-371, 1954.

Zollinger, R. M. and Ellison, E. H.: Primary peptic ulcera-
tions of the jejunum associated with islet cell tumors of the
pancreas. Ann. Surg. 142:709-728, 1955.

Lulu, D. J.; Corcoran, T. E. and Andre, M.: Familial endo-

31.

52.

53.

54.

crine adenomatosis with associated Zollinger-Ellison syn-
drome. Wermer’s syndrome. Amer. J. Surg. 115:695-701,
1968.

Morton, N. E.: The detection and estimation of linkage
between the genes for elliptocytosis and the Rh blood type.
Amer. J. hum. Genet. 8:80-96, 1956.

Senior, B. and Loridan, L.: Functional differentiation of
glycogenoses of the liver with respect to the use of glycerol.
New Engl. J. Med. 279:965-970, 1968.

Mayr, E.: Theory of biological classification. Nature 220:
545-548, 1968.

Temtamy, S. and McKusick, V. A.: The Genetics of Hand
Anomalies. Johns Hopkins Press, Baltimore, 1969.