HIGHLIGHTED ARTICLE

GENETICS | INVESTIGATION

An Estimate of the Average Number of Recessive

Lethal Mutations Carried by Humans
Ziyue Gao,*,1 Darrel Waggoner,†,‡ Matthew Stephens,†,§ Carole Ober,†,** and Molly Przeworski††,‡‡,1
*Committee on Genetics, Genomics, and Systems Biology, †Department of Human Genetics, ‡Department of Pediatrics,
§Department of Statistics, and **Department of Obstetrics and Gynecology, University of Chicago, Chicago, Illinois 60637,
††Department of Biological Sciences, Columbia University, New York, New York 10027, and ‡‡Department of Systems Biology,
Columbia University, New York, New York 10032

ABSTRACT The effects of inbreeding on human health depend critically on the number and severity of recessive, deleterious mutations
carried by individuals. In humans, existing estimates of these quantities are based on comparisons between consanguineous and
nonconsanguineous couples, an approach that confounds socioeconomic and genetic effects of inbreeding. To overcome this
limitation, we focused on a founder population that practices a communal lifestyle, for which there is almost complete Mendelian
disease ascertainment and a known pedigree. Focusing on recessive lethal diseases and simulating allele transmissions, we estimated
that each haploid set of human autosomes carries on average 0.29 (95% credible interval [0.10, 0.84]) recessive alleles that lead to
complete sterility or death by reproductive age when homozygous. Comparison to existing estimates in humans suggests that
a substantial fraction of the total burden imposed by recessive deleterious variants is due to single mutations that lead to sterility or
death between birth and reproductive age. In turn, comparison to estimates from other eukaryotes points to a surprising constancy of
the average number of recessive lethal mutations across organisms with markedly different genome sizes.
KEYWORDS autosomal recessive disease; consanguinity; inbreeding; recessive lethal mutation; human

I N diploid organisms such as humans, the efficacy of selec-

tion on a deleterious mutation depends both on the fitness
of homozygotes and on the fitness of heterozygotes, which
alleles from one or more common ancestors, the genomes of
offspring of consanguineous couples are more likely to be
identical by descent, revealing recessive, deleterious traits. If
reflects dominance relationships among alleles. Since re- there are many recessive or partially recessive deleterious
cently introduced mutations are mostly present in heterozy- mutations (i.e., mutations for which the fitness of the het-
gotes, they will be purged less effectively by selection when erozygote is closer to that of the fitter homozygote) segre-
recessive and segregate at higher frequencies compared to gating in the population, inbred individuals will, on average,
dominant or semidominant alleles that cause a similar fitness have lower fitness than outbred individuals. A reduction in
reduction in homozygotes. For this reason, recessive alleles mean fitness due to inbreeding (“inbreeding depression”)
are expected to constitute a large fraction of strongly dele- has been demonstrated repeatedly in experimental studies
terious alleles segregating in diploid populations and in par- in multiple Drosophila species (e.g., Greenberg and Crow
ticular of Mendelian disease mutations in humans. 1960; reviewed in Simmons and Crow 1977), as well as
One context in which the effects of recessive mutations under seminatural conditions in mice (Meagher et al. 2000).
are unmasked is in the presence of inbreeding, which leads Estimating the burden of recessive deleterious mutations
to an excess of homozygotes compared to Hardy–Weinberg in humans is therefore key to predicting adverse outcomes
expectation. Because closely related individuals may co-inherit of consanguineous unions due to genetic factors (Morton
et al. 1956; Bittles and Makov 1988; Bittles and Neel
Copyright © 2015 by the Genetics Society of America 1994; Bittles and Black 2010a). Two main methods have
doi: 10.1534/genetics.114.173351 been developed to these ends: both aim to quantify the
Manuscript received December 9, 2014; accepted for publication February 11, 2015;
published Early Online February 18, 2015. burden by comparing the health states of offspring of non-
Available freely online through the author-supported open access option.
1
consanguineous and consanguineous matings. The first con-
Corresponding authors: University of Chicago, 920 E. 58th St., CLSC 413, Chicago, IL
60637. E-mail: ziyuegao@uchicago.edu; Columbia University, 602A Fairchild Center,
siders couples with variable degrees of relatedness and
M.C. 2424, New York, NY 10027. E-mail: mp3284@columbia.edu regresses the viabilities of their offspring on their inbreeding

Genetics, Vol. 199, 1243–1254 April 2015 1243

coefficients, F (Morton et al. 1956). When applied to occurs as homozygous in at least one individual) in the past
humans, this method suffers from a number of limitations. few generations, by simulating its transmission down the
For one, the estimate relies heavily on accurate assessments pedigree. From this estimate and the number of recessive
of degrees of relatedness, and yet the F values estimated diseases observed in the pedigree, we can obtain an esti-
from recent pedigrees do not capture inbreeding among mate of the total number of deleterious mutations carried
more distant ancestors. This will bias the results if, as seems by the founders. Since the number of founders is known,
plausible, consanguineous marriages tend to occur in fami- this estimate translates into the average number of recessive
lies with a tradition of close-kin unions (Hussain and Bittles lethal alleles in each haploid set of autosomes. An advantage
2000; Hamamy et al. 2011). Even if F is calculated based on of our approach is that, by directly utilizing the pedigree in-
deeper pedigrees, it represents only the expected proportion formation, there is no need to calculate an inbreeding coeffi-
of the genome that is identical by descent, whereas the re- cient or to compare among groups that are potentially subject
alized proportion could vary tremendously across individu- to different socioeconomic conditions. A difficulty, however,
als. In practice, this variation, combined with sampling is that the transmission probability of a recessive deleterious
variance, can lead to considerable uncertainty in the esti- allele depends on its selection coefficient in homozygotes
mated effects of recessive alleles (Bittles and Makov 1985, (s), which is in general very hard to quantify. We therefore
1988). Moreover, due to the restricted range of F and the focused on autosomal, recessive lethal mutations (s = 1),
small number of data points, the estimate of the combined defined as mutations that when homozygous lead to com-
effect of recessive deleterious mutations is highly sensitive plete sterility or death between birth and reproductive age,
to the choice of the regression model (Makov and Bittles in the absence of medical treatment. Since recessive lethal
1986). Perhaps most importantly, consanguineous and non- mutations are only a subset of all deleterious mutations, our
consanguineous groups differ with respect to socioeconomic estimate provides a lower bound on the burden of recessive
factors, in ways that influence the mortality and morbidity of deleterious mutations, as well as information on the tail of
the progeny (Schull and Neel 1965; Neel et al. 1970; Hussain the distribution of fitness effects of deleterious mutations
and Bittles 1998, 2000). How estimates of genetic effects will (e.g., Yampolsky et al. 2005; Eyre-Walker et al. 2006; Boyko
be affected is unclear, as the strength of the correlation be- et al. 2008).
tween socioeconomic status and inbreeding—and even the
direction of the correlation—varies across societies (e.g.,
Neel et al. 1970). Thus, this approach could either overesti- Materials and Methods
mate or underestimate the genetic effects of consanguinity
Simulations
on health outcomes.
To minimize these concerns, a second approach focuses To assess the probability of a recessive lethal mutation
specifically on the comparison between offspring of first- manifesting itself after 1950, we ran two sets of gene-dropping
cousin marriages and of nonconsanguineous marriages in simulations.
a large number of populations (Bittles and Makov 1988;
Bittles and Neel 1994; Bittles and Black 2010b). Regression Simulations on the minimum pedigree: We first considered
of the mortality of first-cousin progeny—for which the F value a “minimum pedigree,” consisting of individuals ancestral to
is 1/16—on that of nonconsanguineous progeny in the same 1642 Hutterites for whom the genotypes at 14 autosomal
population reveals a significant excess mortality in the former, recessive disease-causing genes were determined previously
which is then translated into the aggregate effect of recessive (Chong et al. 2012). When using this pedigree, we assumed
deleterious mutations. Even in this approach, however, ge- no transmission distortion (Meyer et al. 2012) and complete
netic effects may be confounded by socioeconomic conditions reproductive compensation; i.e., we assumed no correlation
that differ between consanguineous and nonconsanguineous between the average number of surviving children and
groups within a population (Bittles and Neel 1994; Hussain whether the parents are carriers of recessive lethals (Ober
and Bittles 1998). et al. 1999). If reproductive compensation is incomplete, our
Here, we introduce an approach that is not confounded simulations are expected to lead to an underestimate of the
by environmental effects, considering a founder population fraction of recessive lethal alleles purged by selection since
that practices a communal lifestyle, with a known pedigree the founding and an overestimate of the probability of man-
and close to complete disease ascertainment over the past ifestation. In each replicate, we assigned a mutation to one
few generations (Hostetler 1974). Founder populations founder and simulated the genotypes of other individuals
have contributed greatly to the identification of Mendelian generation by generation. For any individual who was
disease mutations, because the founding bottleneck and sub- known to have children, if both his (or her) parents were
sequent inbreeding increase the chance of recent identity- carriers of a recessive lethal, we assigned a heterozygous
by-descent and thus the incidence of a number of otherwise genotype with probability 2/3 and a homozygous beneficial
rare, recessive diseases (Boycott et al. 2008). With a known genotype with probability 1/3. For all other situations, we
pedigree, we can estimate the probability that an autosomal simulated the offspring genotype based on the parental
recessive, deleterious founder mutation manifests itself (i.e., genotypes according to Mendelian inheritance rules. This

1244 Z. Gao et al.

simulation scheme relies on the assumption of complete re- (8.19%) cases. The proportion of recessive lethal mutations
productive compensation, but is robust to missing data in lost is higher than that of neutral variants (60.0%), suggest-
the pedigree. After generating genotypes of all individuals in ing that in addition to the dominant effect of genetic drift,
the pedigree, we examined the numbers of heterozygotes selection could also have played a role in purging out re-
and homozygotes for that mutation among people born after cessive lethals in Hutterites. We also ran 39,000 replicates
1950. The mutation was classified as “lost” if there was no where we considered individuals born after 1940 as the co-
heterozygote or homozygote in the cohort and as “mani- hort (500 replicates for each founder). Among the 34,001
fested” if at least one homozygote was present in the cohort. replicates retained, the mutation was lost in 23,242 (68.4%)
Because individuals in the pedigree are related to each cases and manifested in 2787 (8.20%) cases.
other, the incidences of a disease in the pedigree are not For both scenarios, we also considered the situation where
independent, so we considered the number of unique dis- more than one copy of the same mutation was present in the
eases instead of the number of affected individuals. We per- founders. In each replicate, we randomly sampled two (or
formed 320,000 gene-dropping simulations (5000 replicates three) founders to be the carriers and simulated the genotypes
for each founder). The mutation was lost before 1950 in of other individuals as described above. A total of 100,000
183,767 (57.4%) cases and manifested in 25,964 (8.11%) simulations were run with the minimum pedigree with two or
cases. We also considered all individuals born after 1940 as three carriers, and the mutation was manifested in 16.7% and
the cohort. Among the 320,000 simulations performed, the 25.5% of the cases, respectively. A total of 10,000 simulations
mutation was lost before 1940 in 182,221 (56.9%) cases were performed with the larger pedigree with two or three
and manifested in 26,328 (8.23%) cases. We focus on this carriers, and the mutation was manifested in 14.8% and 20.6%
simulation scheme in the main text. of the cases retained, respectively. These findings indicate
that the probability of manifestation is approximately pro-
Simulations on a larger pedigree: We also considered portional to the number of carriers among the founders,
a larger pedigree, which consists of 15,236 individuals, all of enabling us to estimate the total number of recessive lethal
whom can be traced back to 78 ancestors (who were not alleles (some of which could be copies of the same muta-
necessarily the “minimum ancestors” defined in Martin tion) carried by the founders by dividing the number of
1970). Individuals who fell into one of the following three distinct diseases by the probability of manifestation with
groups were included in this pedigree: (1) before the sepa- one carrier.
ration of the three leuts, individuals who had descendants in
Identification of recessive lethal diseases in Hutterites in
the Schmiedeleut (S-leut); (2) all S-leut individuals who
the pedigree
were born since the separation of the three leuts through
1980; and (3) S-leut individuals who were born between Most Mendelian diseases reported in Hutterites were summa-
1980 and 2013 and participated in our ongoing studies rized in a review by Boycott et al. (2008), and the list of re-
(Chong et al. 2012). When using this larger pedigree, we cessive diseases was further updated by Chong et al. (2012).
assumed no reproductive compensation and no transmission To incorporate newly identified diseases, we searched PubMed
distortion (Meyer et al. 2012). Specifically, in each replicate, for genetic diseases in Hutterites that were reported since
we assigned a mutation to one founder and then simulated 2012. We also searched for diseases with terms “recessive”
the genotypes of all other individuals according to Mende- and “Hutterites” in the Online Mendelian Inheritance in Man
lian inheritance rules. Because individuals homozygous for (OMIM) (an online catalog of human genetic disorders and
a recessive lethal mutation cannot reproduce, any individual underlying genes) and confirmed that all entries are included
who has offspring cannot be a homozygote recessive. To in our list.
model this property of recessive lethals, we retained only We then classified a disease as “recessive lethal” if (i) it
replicates that were consistent with this condition, i.e., rep- has 100% penetrance in homozygotes, (ii) the heterozygotes
licates where all individuals with offspring were either het- are asymptomatic (although they may have subtly decreased
erozygous or homozygous for the beneficial allele. This fitness), and (iii) the disease leads to prereproductive lethal-
simulation scheme would be exact on a complete pedigree, ity (e.g., restrictive dermopathy, cystic fibrosis in females) or
in which differences in family size are entirely attributable to complete sterility (e.g., cystic fibrosis in males) in the ab-
recessive lethals and stochasticity. However, it is sensitive to sence of treatment. We included infertility due to biological
incompleteness of the pedigree and other nonrandom fac- reasons (e.g., cystic fibrosis in males) or inability to repro-
tors that affect family size, in ways that are expected to lead duce if the phenotype of the condition leads to a reproduc-
to an overestimate of the fraction of recessive lethal alleles tive fitness of zero due to social barriers (e.g., nonsyndromic
purged by selection since the founding and hence an un- mental retardation, myopathy with movement disorder and
derestimate of the probability of manifestation. We per- intellectual disability).
formed 78,000 gene-dropping simulations on the larger To restrict the number of recessive lethal diseases to the
pedigree (1000 replicates for each founder). Among the pedigree under study, we required there to be affected
67,964 replicates retained, the mutation was lost before individuals in Hutterites in South Dakota. This narrowed
1950 in 46,719 (68.7%) cases and manifested in 5563 down the list to four diseases (cystic fibrosis, nonsyndromic

Mean Number of Recessive Lethals 1245

mental retardation, restrictive dermopathy, and myopathy Yi jXi  BinðXi ; pi Þ: (3)
with movement disorder and intellectual disability). We ex-
cluded restrictive dermopathy when considering the minimum Because of the thinning property of Poisson distribution,
pedigree, because the only reported patient in S-leut was not conditional on R, Yi follows a Poisson distribution:
included in the minimum pedigree (although the parents
Yi jR ¼ r  Poissonð2rpi Þ:
were included and confirmed to be carriers by genotyping)
(Chong et al. 2012; Loucks et al. 2012). For the other three
diseases, genotype data of the 1642 extant individuals con- Moreover, because each Yi depends only on the corre-
firmed the presence of an individual(s) homozygous for the sponding Xi and, conditional on R, the Xi’s are independent
disease allele in the pedigree (Chong et al. 2012). of each other, the Yi’s are independent Poisson random vari-
We note that two CFTR mutations have been identified in ables. Therefore, the total number of diseases observed, D =
the Hutterites. Both alleles lead to severe phenotypes such Y1 + Y2 + . . . + YNf, is a Poisson random variable
that homozygous or compound heterozygous males are
completely sterile and homozygous or compound heterozy- DjR ¼ r  Poissonð2rpNf Þ; (4)
gous females cannot survive to reproductive age in absence
where P = (p1 + p2 + . . . + pNf)/Nf represents the proba-
of treatment. We therefore treat them as two copies of the
bility of manifestation if each mutation was equally likely to
same recessive lethal mutation. We further note that although
be carried by each founder, which can be estimated from our
the p.F508del mutation is common in Europeans, it is present
simulations.
on a single haplotype in Hutterites, suggesting that it was in-
Combining (1) and (4), we can rewrite the posterior
troduced into the population by only one founder (Chong et al.
distribution of R as
2012). The p.M1101K mutation is rare in Europeans but was
identified on two haplotypes in Hutterites (Zielenski et al.   
fR ðrÞ ð2rpNf ÞD D! e22rpNf
1993). The two haplotypes differ at multiple sites on both sides fRjD ðrÞ ¼ R  
of the mutation, indicating either that at least two recombina- N D 22rpNf dr
r¼0 fR ðrÞ ð2rpNf Þ D! e
tion events have occurred or that the p.M1101K mutation was
introduced by two founders (Chong et al. 2012). Therefore, it ð2rpNf ÞD 22rpNf
} fR ðrÞ e :
is likely that two or three carriers of these two CFTR mutations D!
were present in the founders. Given that the probability of
manifesting a mutation is approximately proportional to the
Because there is little definitive information about R, we
number of carriers in the founders, we can treat it as intro-
use a uniform prior on (0, N),
duced by only one founder.
Point estimation and credible intervals for the mean fR ðrÞ ¼ 1;
number of mutations per haploid human genome
as well as a uniform prior of the logarithm of R on (0, N),
We used a Bayesian approach to estimate the credible
interval for mean number of recessive lethal alleles carried 1
fR ðrÞ ¼ :
by each haploid set of human autosomes, R. Given that D r
recessive lethal diseases have been observed, the posterior
probability of R is While both are improper priors, the resulting posterior
distributions are proper gamma distributions:
fR ðrÞPðD j R ¼ rÞ 
fRjD ðrÞ ¼ R N } fR ðrÞPðDR ¼ rÞ; (1)
r¼0 fR ðrÞPðD j R ¼ rÞdr
fRjD ðrÞ  GammaðD þ 1; 2pNf Þ; if a uniform prior;

where fR|D(r) is the conditional probability of observing D or

diseases, and fR(r) is the prior on R.
Let Xi be the number of unique recessive lethal mutations fRjD ðrÞ  GammaðD; 2pNf Þ; if a uniform prior on the
carried by the ith founder, among which Yi mutations man- logarithmic scale:
ifested themselves in the pedigree. We assume that each Xi is
independently Poisson distributed:
Therefore, the posterior mode is R = D/2pNf, if we as-
Xi jR ¼ r  Poissonð2rÞ: (2) sume a uniform prior [or R = (D 2 1)/2pNf, if a uniform
prior on the logarithmic scale]. Given the value of p
For simplicity, we assume all mutations carried by the obtained from the simulations, the 95% credible interval
founders are unique. If the transmission of each mutation is (CI) of the posterior distribution was determined by using
independent and each of the Xi mutations carried by the ith R (R Development Core Team 2013). For simulations with
founder has a manifestation probability of pi, then Yi follows the minimum pedigree (D = 3, P = 0.0811, Nf = 64), the CI
a binomial distribution: for R is [0.10, 0.84] when a uniform prior is assumed and

1246 Z. Gao et al.

[0.060, 0.70] with a uniform prior on the logarithmic scale. f1i pffiffiffiffiffiffiffiffiffiffiffiffi
Similar results are obtained from the larger pedigree (D = 4, qi ¼ ¼ mi 2pNe :
2
P = 0.0819, Nf = 78): the CI for R is [0.13, 0.80] with
a uniform prior and [0.085, 0.69] with a uniform prior on Therefore, the average number of recessive lethal mutations
the logarithmic scale. carried by a haploid genome is

The target size for mutations that lead to recessive X

m X
m pffiffiffiffiffiffiffiffiffiffiffiffi X
m pffiffiffiffiffiffiffiffiffiffiffiffi
qi ¼ mi 2pNe ¼ li mbp 2pNe
lethal disorders between birth and reproductive age
i¼1 i¼1 i¼1
A genomic parameter of interest is the total number of pffiffiffiffiffiffiffiffiffiffiffiffi X
m

functionally important sites that, if mutated, would give rise ¼ mbp 2pNe li ;
i¼1
to recessive lethal alleles. We term this quantity the “target Pm
size” for recessive lethal disorders and infer it from our es- where i¼1 li is the target size for all autosomal recessive
timate of R. lethal mutations.
Assuming that the founders of S-leut Hutterites were
The completely recessive case: Using a diffusion model and drawn from a random-mating population at equilibrium,
a low mutation rate approximation, Li and Nei (1972) de- each of them should have carried twice that number of
rived the expectation for the total number of heterozygotes recessive lethals. Based on our estimate of R = 0.29, a mu-
that carry a recessive deleterious mutation in a finite pop- tation rate of 1.2 3 1028/bp per generation (Campbell et al.
ulation [designated by n1(p)], 2012), and a diploid effective population size of 20,000, the
rffiffiffiffiffiffiffiffiffiffi target size is then estimated to be 6.8 3 104 bp.
p In a population with larger Ne, the efficacy of purifying
n1 ðpÞ ¼ 4NNe p ; for Ne s $ 5
2Ne s selection against recessive alleles increases (proportional to
Ne 1=2 ), but the increase in the mutational input (proportional
(equation 7b in Li and Nei 1972), where p is the initial to Ne) is greater than the increase in the efficacy of selection
frequency of the mutation, N and Ne the actual and effective (Simons et al. 2014). Therefore, at equilibrium, an individual
population sizes, respectively, and s the selection coefficient in a population with larger Ne will have a greater expected
against homozygotes (Li and Nei 1972). number of recessive lethal mutations. The recent population
The total number of heterozygotes affected by a single growth experienced by humans represents a transition from
recessive lethal mutation, n1 ; can be obtained by replacing p small Ne to large Ne, which will therefore lead to an increase
and s by 1/(2N) and 1 in the above equation: in the average number of recessive lethal mutations per indi-
pffiffiffiffiffiffiffiffiffiffiffiffi vidual. As a result, the estimated target size after recent growth
n1 ¼ 2pNe : will be smaller than that estimated from the long-term Ne.

For a gene with a mutation rate to recessive deleterious The partially recessive case: If a deleterious mutation leads
alleles of m; the expected frequency of heterozygotes in to complete lethality (or sterility) in homozygotes (s = 1)
a random generation is then and a decrease of hs in fitness of heterozygotes, selection
against the deleterious allele would mainly come from the
f1 ¼ 2mn1 death of heterozygotes, because the death of homozygotes is
a rare event (when the allele frequency is low). As in the
(equation 18 in Li and Nei 1972). For simplicity, we assume completely recessive case, we assume that there are m au-
that there are m autosomal genes in the genome that each tosomal genes in the genome that can mutate to deleterious
can lead to complete sterility or lethality between birth and alleles with these properties and that gene i has li such sites.
reproductive age and that gene i has li sites at which muta- Under these assumptions, the expected total number of het-
tions will give rise to recessive lethal alleles. We further erozygotes affected by a mutation in a finite population at
assume that each site has the same per generation mutation gene i can be approximated as
rate, mbp, so the total mutation rate to recessive lethal alleles
1 1
at gene i is n1 ¼ ¼ ; for 4Ne sh $ 1
hs h
mi ¼ li mbp :
(equation 6b in Li and Nei 1972). So the average frequency
The expected frequency of heterozygotes carrying a recessive of those partially recessive lethal mutations is
lethal allele at this gene i is mi
qi ¼ mi n1 ¼ ;
pffiffiffiffiffiffiffiffiffiffiffiffi h
f1i ¼ 2mi n1 ¼ 2mi 2pNe ;
which is independent of the effective population size and the
and the expected frequency of recessive lethal alleles at this same as obtained assuming mutation–selection balance
gene is approximately (Haldane 1935).

Mean Number of Recessive Lethals 1247

 Therefore, the total number of such partially recessive to complete—and inherited two copies of a recessive lethal
lethal mutations carried by a random haploid genome is variant introduced by the founder.
We found that on average 57% of unique recessive lethal
X
m X
m
m X
m lm
i bp mbp X
m
mutations carried by the founders have been lost before
i
qi ¼ ¼ ¼ li ;
i¼1 i¼1
h i¼1
h h i¼1 1950 (i.e., none of the individuals in the pedigree born after
P 1950 carries the mutation). This proportion is almost the
where m i¼1 li is the target size for all mutations that lead to same as for neutral variants with the same initial frequency
lethality in homozygotes and a fitness decrease of hs in het- in the founders, suggesting that the loss of variants is pri-
erozygotes. Plugging in s = 1, h = 0.01 and taking a muta- marily due to the severe genetic drift after the founder event
tion rate of 1.2 3 1028/bp per generation (Campbell et al. (confirming results found in Chong et al. 2012). Among the
2012), we estimate that the target size of such nearly re- recessive lethal mutations that survive until 1950, we expect
cessive mutations is 2.4 3 105 bp. that 19% will have been exposed in homozogote(s) and
thus, in expectation, 8.1% of all recessive lethal mutations
carried by the founders will have manifested themselves to
Results
date. The probability is almost the same if we consider indi-
We focused on the Hutterites, a group ideally suited for the viduals born after 1940 as the cohort, to allow for a delay in
new method we propose. The Hutterian Brethren is an diagnosis for diseases with an onset in adolescence (see
isolated founder population, which originated in the Tyro- Materials and Methods).
lean Alps in the 1520s and eventually settled in North The above simulation scheme implicitly assumes com-
America on three communal farms in the 1870s after a series plete reproductive compensation (see Materials and Meth-
of migrations throughout Europe. The three colonies thrived ods), which might not be appropriate for all recessive
and shortly thereafter gave rise to three major subdivisions, lethal diseases. To address this concern, we ran a second
referred to as the Schmiedeleut (S-leut), Lehrerleut (L-leut) set of simulations on a larger pedigree of 15,235 Hutterites,
and Dariusleut (D-leut), with most marriages since 1910 who can be traced back to 78 ancestors (see Materials and
taking place among individuals within the same leut. Each Methods for details about this pedigree). The second simu-
leut practices a communal lifestyle, with no private owner- lation scheme assumes no reproductive compensation and
ship and hence no socioeconomic differences among mem- would be exact on a complete pedigree in which differences
bers (Hostetler 1974). in family size are entirely attributable to recessive lethals
The Hutterites have kept extensive genealogical records, and stochasticity, but it is sensitive to the incompleteness
from which highly reliable pedigrees have been recon- of the pedigree (especially missing data in the early gener-
structed (Bleibtreu 1964; Mange 1964; Steinberg et al. ations). Nonetheless, the results are similar: on average,
1967). Moreover, researchers and the Hutterites community 69% of recessive lethal mutations carried by the founders
have built a close partnership, greatly facilitating the diag- will have been lost before 1950 and 8.2% will have mani-
nosis and identification of genetic disorders. Incidences of fested themselves in individuals born after 1950.
genetic disorders in Hutterites have been comprehensively Next, we considered all known autosomal recessive lethal
documented since the late 1950s, with .40 Mendelian dis- diseases observed in the S-leut Hutterites included in the
orders, of which 35 are autosomal recessive, described in the pedigree. To this end, we compiled a list of all known au-
literature (Boycott et al. 2008, 2010; Çalışkan et al. 2011; tosomal recessive disorders (11 in total) reported in S-leut
Huang et al. 2011; Bögershausen et al. 2013; Gerull et al. Hutterites in the United States (see Materials and Methods
2013; Wiltshire et al. 2013). The pedigree information and for details). We classified each of the identified diseases as
nearly complete ascertainment of genetic disorders over the “lethal” if, untreated, it leads to death between birth and
past 60 years make it possible to reliably infer the number reproductive age or precludes reproduction for affected indi-
of recessive lethal mutations in the founders. viduals. We found four such recessive lethal diseases: cystic
Specifically, our analysis was based on a 13-generation fibrosis, nonsyndromic mental retardation, a severe form of
pedigree that relates 1642 extant S-leut Hutterites in South myopathy, and restrictive dermopathy (Table 1). The under-
Dakota and their ancestors (3657 individuals in total), all of lying mutations are known for all four diseases, and geno-
whom can be traced back to 64 founders who lived in the typing data of the 1642 extant individuals confirmed the
early 18th to early 19th centuries in Europe (Chong et al. presence of homozygote(s) of the disease-causing muta-
2012). We ran gene-dropping simulations (similar to those tions for the first three diseases (Chong et al. 2012). Re-
in Manatrinon et al. 2009 and Chong et al. 2012) on the strictive dermopathy was excluded from our list when we
pedigree to assess the probability of an autosomal recessive used the 13-generation pedigree, because the only known
lethal mutation manifesting itself, conditional on being car- case among the South Dakota Hutterites was not included
ried by a founder (see Materials and Methods for simulation in the 3657 individuals.
procedures). We defined “manifestation” as the presence of From the number of recessive lethal diseases (three) and
at least one individual in the pedigree who was born after the probability that a recessive lethal allele manifested itself
1950—the period after which disease ascertainment is close since 1950 (0.081), we estimated that the total number of

1248 Z. Gao et al.

Table 1 Three autosomal recessive lethal diseases and corresponding mutations observed in the smaller pedigree

OMIM Carrier frequency Carrier frequency in

Name of disease phenotype no. Gene Mutation in Europeans S-leut Hutterites (%) Source
Cystic fibrosisa 219700 CFTR p.F508dela 2.2–3.9% 2.2b Fujiwara et al.
p.M1101Ka Unknown (one 7.3b (1989); Zielenski
case reported) et al. (1993)
Nonsyndromic mental 614020 TECR p.P182L Found only in 6.9b Çalısxkan
retardation Hutterites to date et al. (2011)
Myopathy with movement — TRAPPC11 c.1287+5G . A Found only in 7 Bögershausen
disorder and intellectual Hutterites to date et al. (2013)
disability
a
See Materials and Methods for treatment of the two mutations in CFTR.
b
The allele frequencies are reported in Chong et al. (2012).

autosomal, recessive lethal mutations carried by the 64 and Europeans despite the out-of-Africa bottleneck. A simi-
founders is 3/0.081 = 37 or an average of 0.29 recessive lar concern might be that a long period of endogamy could
lethal alleles in each haploid human genome (Figure 1A). have purged recessive deleterious alleles from the popula-
To assess the uncertainty in this estimate, we estimated the tion that led to the Hutterites (Keller and Waller 2002).
posterior distribution of the mean number of mutations per There is no evidence that such a demographic scenario oc-
haploid human genome conditional on observing exactly curred, but if it did, it is again unlikely to have had much of
three diseases since 1950 (Materials and Methods and Figure an effect: over the 15 generations between the origin of
1B). If a uniform prior distribution is used, the posterior the Hutterites in the 1520s and the founding event, even
distribution has a mode of 0.29 and a 95% CI of [0.10, relatively high levels of human inbreeding (F = 0.03) should
0.84]. We also considered a uniform prior on the logarith- decrease the mean allele frequency of recessive lethals only
mic scale to allow for uncertainty in the order of magnitude, by 30% (Overall et al. 2002). Moreover, such a decrease
and a similar 95% CI is obtained (i.e., [0.060, 0.70] muta- would be lessened or nullified by reproductive compensa-
tions per haploid genome). The point estimate and 95% CI tion (Overall et al. 2002), as might occur in the Hutterites
were similar, when we used simulation results from the (Ober et al. 1999). These considerations suggest that esti-
larger pedigree (see Materials and Methods). mates based on the Hutterites should be broadly applicable
Simulations further indicate that only a small fraction of and would, if anything, be slightly lower than the mean
the surviving recessive lethal mutations have been seen in number of recessive lethals carried by larger, outbred
homozygotes, so there are more hidden, recessive lethal populations.
mutations that are segregating among extant individuals in In that regard, we note that our estimate of the average
the pedigree. In fact, carrier screening has identified number of recessive lethal mutations per haploid genome is
heterozygotes for three more recessive lethal mutations in lower than the previous estimates of the total number of
the S-leut Hutterites in South Dakota, which have man- “lethal equivalents” per haploid genome (0.56–0.7 in Bittles
ifested themselves in Hutterites outside the pedigree under and Neel 1994; Bittles and Black 2010b). A lethal equivalent
study (Table 2) (Chong et al. 2012). Based on our simula- is defined as a locus or a set of loci that, when in the homo-
tion results, we expect quite a few more recessive lethal zygous state, would cause on average one death, e.g., one
mutations in addition to these cases, most of which remain lethal mutation or two mutations each with 50% probability
unknown. of causing death (Morton et al. 1956). In other words, the
In generalizing from the results for the Hutterites to other total number of lethal equivalents in a haploid genome can
human populations, one concern might be that their de- be thought of as the sum of the deleterious effects of all
mographic history prior to the founder event in the 18th– recessive mutations carried by an individual. Comparison
19th centuries was atypical in ways that influence the num- to estimates of this quantity suggests that, as expected, re-
ber of recessive lethals carried by the founders. While tran- cessive lethal mutations are only a subset of the recessive
sient demographic changes can have a marked impact on mutational burden. Interestingly, however, the difference
patterns of genetic variation, they are not expected to have between our point estimate and previous estimates is only
a substantial effect on the average number of recessive le- about twofold; even if we consider the lower bound of our
thal alleles carried by an individual, because their equilib- credible interval on the mean number of recessive lethals, it
rium frequency is reestablished on a relatively short is still about one-sixth of the total number of lethal equiv-
timescale (Balick et al. 2014). For instance, after a bottle- alents. Thus, insofar as previous estimates are reliable, it
neck, this quantity of interest returns to the equilibrium appears that a substantial portion of the total burden of re-
pffiffiffiffiffiffiffiffiffi
value within 4N0 generations (where N0 is the original cessive mutations carried by humans is attributable to single
population size before the bottleneck); for this reason, this mutations that, when homozygous, lead to sterility or death
quantity is expected to be very similar for modern Africans between birth and reproductive age.

Mean Number of Recessive Lethals 1249

 Figure 1 The analysis pipeline for estimating the
average number of recessive lethal mutations car-
ried by humans. (A) A schematic diagram of the
approach. The analysis procedures are described in
black. Specific values and estimates for the Hutter-
ites are provided in blue, including the point esti-
mate of the mean number of mutations carried by
a founder. (B) The posterior distribution of the
mean number of recessive lethal mutations carried
by each haploid human genome, given the prob-
ability of manifestation and the number of diseases
observed.

Discussion simulating the transmission of those mutations, this should not

influence the mean proportion of mutations that survive or
Our approach indicates that on average, each Hutterite
manifest themselves, so will not bias our estimate of the prob-
founder carried 0.58 autosomal, recessive lethal mutations
ability of manifestation. We also ignore the possibility that de
that lead to death between birth and reproductive age or to
novo mutations that arose since the founding may contribute to
complete sterility. This is likely a slight underestimate for other the diseases considered. This assumption is justified because
human populations, if we take into account the effects of the we expect at most a few recessive lethal mutations to have
unique demographic history of the Hutterites. Nonetheless, this arisen in the pedigree in the 13 generations (see discussion
estimate is unaffected by socioeconomic factors. Moreover, in- on the target size of recessive lethals below) and their proba-
complete ascertainment of diseases is unlikely to be a major bility of manifestation is even lower than that of founder muta-
concern, because most severe genetic disorders that occurred tions. Thus, overall, we expect the estimate to be relatively
in Hutterites after the 1950s are expected to have been docu- unbiased and, if anything, a slight underestimate of the mean
mented (Boycott et al. 2008), so we expect at most a slight value of other populations.
underestimate of their number. In addition, while we ignore If we take our point estimate at face value, it suggests
interference and linkage between recessive lethal alleles when that the risk of autosomal recessive lethal disorders that

1250 Z. Gao et al.

Table 2 Three autosomal recessive lethal mutations found only in heterozygotes in the pedigree of S-leut Hutterites in South Dakota

OMIM Carrier frequency Carrier frequency

Name of disease phenotype no. Gene Mutation in Europeans in S-leut Hutterites Source
Bardet–Biedl syndrome 209900 BBS2 c.472-2A . G Found only in Hutterites 2.8a Innes et al. (2010)
to date
Dilated cardiomyopathy 610198 DNAJC19 IVS3-1G . C Found only in Hutterites 2.8a Davey et al. (2006)
with ataxia syndrome to date
Joubert syndrome/Meckel 614424 TMEM237 p.R18X Found only in Hutterites 8.0a Schurig et al. (1980);
syndrome to date Huang et al. (2011)
a
The allele frequencies are reported in Chong et al. (2012).

manifest after birth should be increased by 0.29/16 = 1.8% sites are of particular interest, because they are of critical
in offspring of first-cousin couples (assuming no difference functional importance; on the other hand, mutations at
in environmental factors). This prediction agrees well with those sites are haplosufficient, in that one functional copy of
the estimated 3.5–4.4% increased risk for prereproductive the gene is enough to maintain fitness. Assuming a random-
mortality and 1.7–2.8% increased incidence of congenital mating, diploid population with constant effective popula-
anomalies in children of first cousins above the general pop- tion size of 20,000 (as a proxy for the population from which
ulation risk (Hamamy et al. 2011). Hutterite founders derived), a mutation rate of 1.2 3 1028/bp
Beyond the Hutterites, this approach can be applied to per generation, and an estimate of 0.29 recessive lethal mu-
other isolated founder populations with limited immigra- tations per haploid set of autosomes, we predict that there
tion, for which there is reliable genealogical information should be 68,000 autosomal base pairs at which mutations
since the founding and close to complete disease phenotyp- lead to recessive lethal disorders on or after birth (see Mate-
ing in the relatively recent past, such as the Amish and the rials and Methods; using theory from Li and Nei 1972).
inhabitants of Norfolk Island (Macgregor et al. 2010; Hinckley Based on this estimate of the target size, we do not expect
et al. 2013). de novo recessive lethals that arose since the founding to
In interpreting our estimates, an important consideration is manifest themselves as diseases since the 1950s. Consider-
that they are limited to lethal diseases that manifest themselves ing a model of population growth that leads the current
after birth. This issue is common to most studies that estimate effective population size to be .20,000, the expected fre-
the mutational burden in humans, because of the limited quency of recessive lethal mutations in the population
availability and reliability of data on prenatal loss. Studies that should be higher (see Materials and Methods), and the esti-
considered data on the frequency of miscarriages (i.e., a gesta- mated size of target size smaller. While these estimates
tion age of $28 weeks) reported no or little effect of consan- should not be taken too literally, as many recessive disease
guinity on prenatal losses, while detecting clear-cut effects on mutations are not point mutations (e.g., Boone et al. 2013),
postnatal mortality (Schull et al. 1970; Bittles and Makov they provide a sense of the number of sites at which reces-
1988; Bittles and Black 2010b). This negative finding cannot sive lethal disease mutations may be discovered.
be taken as strong evidence for the absence of embryonic re- Moreover, this estimate of target size provides comple-
cessive lethals in humans, as most losses due to embryonic mentary information to population genetic approaches that
lethals may occur during earlier stages of pregnancy. Even if aim to estimate the distribution of fitness effects of muta-
the data on early pregnancy loss were available, the high rate tions from polymorphism and divergence and mostly learn
of spontaneous pregnancy failure due to other causes (Leridon about weaker selection coefficients (Eyre-Walker and Keightley
1977) may obscure the difference between consanguineous 2007). These methods find that 25–40% of all amino acid
and nonconsanguineous groups due to embryonic recessive changes in humans are strongly deleterious (i.e., have s .
lethals. In contrast to how little is known in humans, extensive 1% in a genic selection model) (Yampolsky et al. 2005; Eyre-
mutation screens in mice reveal a high proportion (40–50%) of Walker et al. 2006; Boyko et al. 2008). Combining these
autosomal knockout mutations that cause deaths in prenatal estimates with our estimated target size would then suggest
stages when homozygous (Mitchell et al. 2001; White et al. that 0.5–0.8% of strongly deleterious mutations are reces-
2013). If the proportion of embryonic lethals is similar for sive lethals that are fatal between birth and reproductive
spontaneous mutations in humans, each human individual car- age.
ries approximately one to two recessive lethal mutations that An important caveat is that recessive disease-causing
act across ontogenesis. mutations may not be completely recessive, in that carriers
of one copy may also have a slight decrease in their fitness
Genomic target size for recessive lethal mutations
that is too subtle to be detectable in clinical diagnosis. If so,
Our results provide insight into the total number of the mutations will segregate in the population at much
autosomal sites in the human genome that, if mutated, give lower frequencies due to selection against heterozygotes,
rise to recessive (or nearly recessive) lethal alleles. Those and the target size could be larger. For instance, if the

Mean Number of Recessive Lethals 1251

heterozygous effect is a 1% decrease in fitness, the corre- frequency of recessive lethal mutations in humans compared
sponding target size would become 240,000 bp (see with more outbred organisms.
Materials and Methods).
Comparison to other species Acknowledgments
Intriguingly, our estimate of the average number of re- We thank the Hutterites for their ongoing support and
cessive lethal mutations per individual is in good agree- participation in our studies. We thank Jessica Chong for helpful
ment with what has been determined experimentally in comments regarding the pedigree data and simulations; Mark
a number of other diploid animal species. Most studies Abney for discussing the simulation procedures; Kym Boycott
were conducted in Drosophila melanogaster, where indi- and Micheil Innes for their help in clarifying which recessive
viduals from natural or laboratory populations were lethal diseases were present in S-leut Hutterites; and Graham
made homozygous to measure the effects on viability. Coop, Dick Hudson, Marty Kreitman, Charles Langley, Noah
The results are relatively consistent, with on average Rosenberg, and Guy Sella for helpful discussions. This study
24.7% of the second chromosomes and 40.7% of the third was supported in part by grants R01 HD21244 and R01
chromosomes in the population carrying at least one HL085197. Z.G. was supported in part by the William Rainey
recessive lethal (or nearly lethal) mutation (Simmons Harper Fellowship from the University of Chicago.
and Crow 1977). Assuming the number of such mutations
is Poisson distributed, this implies that each D. melanogaster
harbors on average 1.6 autosomal recessive lethal muta- Literature Cited
tions. Similar numbers were obtained in other Drosophila
species (e.g., Dobzhansky et al. 1963; Malogolowkin-Cohen Alexander, R. P., G. Fang, J. Rozowsky, M. Snyder, and M. B. Gerstein,
et al. 1964), as well as by sib crosses in Lucania goodie (1.87 2010 Annotating non-coding regions of the genome. Nat. Rev.
Genet. 11: 559–571.
lethal mutations per individual) and Danio rerio (1.43 muta-
Balick, D. J., R. Do, D. Reich and S. Sunyeav, 2014 Response to
tions) (McCune et al. 2002). a population bottleneck can be used to infer recessive selection.
Our estimate in humans (of approximately one to two arXiv: 1312.3028.
recessive lethals that act across all developmental stages per Bittles, A. H., and M. L. Black, 2010a Consanguinity, human evo-
diploid) is again quite similar, reviving the question of why lution, and complex diseases. Proc. Natl. Acad. Sci. USA 107:
such distantly related organisms carry similar numbers of 1779–1786.
Bittles, A. H., and M. L. Black, 2010b The impact of consanguinity
recessive lethal mutations per genome, despite their highly
on neonatal and infant health. Early Hum. Dev. 86: 737–741.
variable genome sizes (McCune et al. 2002; Halligan and Bittles, A. H., and U. E. Makov, 1985 Linear regressions in the
Keightley 2003). Under a model of mutation–selection bal- calculation of lethal gene equivalents in man. Ann. Hum. Biol.
ance, the equilibrium frequency of a recessive lethal allele in 12: 287–289.
an outbred population depends solely on the mutation rate Bittles, A. H., and E. Makov, 1988 Inbreeding in human popula-
(Gillespie 2004). Given that the mutation rates per base pair tions: an assessment of the costs, pp. 153–167 in Human Mating
Patterns, edited by C. G. N. Mascie-Taylor, and A. J. Boyce. Cam-
per generation are thought to be similar across vertebrates
bridge University Press, Cambridge, UK/New York.
and Drosophila (Ségurel et al. 2014), organisms with larger Bittles, A. H., and J. V. Neel, 1994 The costs of human inbreeding
genome sizes are expected to carry more recessive lethals. and their implications for variations at the DNA level. Nat.
This seemingly surprising constancy is reminiscent of the Genet. 8: 117–121.
C-value paradox—the observation that the genome size of Bleibtreu, H., 1964 Marriage and Residence Patterns in a Genetic
eukaryotes appears to reflect neither the organismal com- Isolate. PhD Thesis, Anthropology Department, Harvard Univer-
sity, Cambridge, MA.
plexity nor the gene number (Thomas 1971). Although both Bögershausen, N., N. Shahrzad, J. X. Chong, J. C. von Kleist-Retzow,
observations suggest that the genome size is a poor predic- D. Stanga et al., 2013 Recessive TRAPPC11 mutations cause
tor of the functional content of a genome, our finding high- a disease spectrum of limb girdle muscular dystrophy and myop-
lights another aspect: the number of sites of crucial athy with movement disorder and intellectual disability. Am. J.
functional importance may be unexpectedly constant across Hum. Genet. 93: 181–190.
taxa, despite the differences in the number of genes and the Boone, P. M., I. M. Campbell, B. C. Baggett, Z. T. Soens, M. M. Rao
et al., 2013 Deletions of recessive disease genes: CNV contri-
length of the coding regions (Alexander et al. 2010; Howe
bution to carrier states and disease-causing alleles. Genome Res.
et al. 2013). Another potential contributing factor to the 23: 1383–1394.
finding in humans could be the mating patterns. Over the Boycott, K. M., J. S. Parboosingh, B. N. Chodirker, R. B. Lowry, D. R.
past centuries, consanguineous marriages were a common McLeod et al., 2008 Clinical genetics and the Hutterite popu-
practice, and the custom remains prevalent in many coun- lation: a review of Mendelian disorders. Am. J. Med. Genet. A.
tries (Bittles and Black 2010a; Hamamy et al. 2011). In the 146A: 1088–1098.
Boycott, K. M., C. Beaulieu, E. G. Puffenberger, D. R. McLeod, J. S.
long term, depending on the degree of reproductive com-
Parboosingh et al., 2010 A novel autosomal recessive malfor-
pensation (Overall et al. 2002), inbreeding can facilitate the mation syndrome associated with developmental delay and dis-
purging of recessive deleterious mutations from the population tinctive facies maps to 16ptel in the Hutterite population. Am. J.
(Keller and Waller 2002), leading to a lower equilibrium mean Med. Genet. A. 152A: 1349–1356.

1252 Z. Gao et al.

Boyko, A. R., S. H. Williamson, A. R. Indap, J. D. Degenhardt, R. D. Hussain, R., and A. H. Bittles, 2000 Sociodemographic correlates
Hernandez et al., 2008 Assessing the evolutionary impact of of consanguineous marriage in the Muslim population of India.
amino acid mutations in the human genome. PLoS Genet. 4: J. Biosoc. Sci. 32: 433–442.
e1000083. Innes, A. M., K. M. Boycott, E. G. Puffenberger, D. Redl, I. M.
Çalışkan, M., J. X. Chong, L. Uricchio, R. Anderson, P. Chen et al., MacDonald et al., 2010 A founder mutation in BBS2 is respon-
2011 Exome sequencing reveals a novel mutation for autosomal sible for Bardet-Biedl syndrome in the Hutterite population:
recessive non-syndromic mental retardation in the TECR gene utility of SNP arrays in genetically heterogeneous disorders.
on chromosome 19p13. Hum. Mol. Genet. 20: 1285–1289. Clin. Genet. 78: 424–431.
Campbell, C. D., J. X. Chong, M. Malig, A. Ko, B. L. Dumont et al., Keller, L., and D. Waller, 2002 Inbreeding effects in wild popula-
2012 Estimating the human mutation rate using autozygosity tions. Trends Ecol. Evol. 17: 230–241.
in a founder population. Nat. Genet. 44: 1277–1281. Leridon, H., 1977 Human Fertility: The Basic Components. Univer-
Chong, J. X., R. Ouwenga, R. L. Anderson, D. J. Waggoner, and C. sity of Chicago Press, Chicago.
Ober, 2012 A population-based study of autosomal-recessive Li, W. H., and M. Nei, 1972 Total number of individuals affected
disease-causing mutations in a founder population. Am. J. Hum. by a single deleterious mutation in a finite population. Am. J.
Genet. 91: 608–620. Hum. Genet. 24: 667–679.
Davey, K. M., J. S. Parboosingh, D. R. McLeod, A. Chan, R. Casey Loucks, C., J. S. Parboosingh, J. X. Chong, C. Ober, V. M. Siu
et al., 2006 Mutation of DNAJC19, a human homologue of et al., 2012 A shared founder mutation underlies restrictive
yeast inner mitochondrial membrane co-chaperones, causes dermopathy in Old Colony (Dutch-German) Mennonite and
DCMA syndrome, a novel autosomal recessive Barth syndrome- Hutterite patients in North America. Am. J. Med. Genet. A.
like condition. J. Med. Genet. 43: 385–393. 158A: 1229–1232.
Dobzhansky, T., B. Spassky, and T. Tidwell, 1963 Genetics of nat- Macgregor, S., C. Bellis, R. A. Lea, H. Cox, T. Dyer et al., 2010 Legacy
ural populations. 32. Inbreeding and the mutational and bal- of mutiny on the Bounty: founder effect and admixture on Nor-
anced genetic loads in natural populations of Drosophila folk Island. Eur. J. Hum. Genet. 18: 67–72.
pseudoobscura. Genetics 48: 361–373. Makov, E., and A. H. Bittles, 1986 On the choice of mathematical
Eyre-Walker, A., and P. D. Keightley, 2007 The distribution of models for the estimation of lethal gene equivalents in man.
fitness effects of new mutations. Nat. Rev. Genet. 8: 610–618. Heredity 57(Pt 3): 377–380.
Eyre-Walker, A., M. Woolfit, and T. Phelps, 2006 The distribution Malogolowkin-Cohen, C., H. Levene, N. P. Dobzhansky, and A. S.
of fitness effects of new deleterious amino acid mutations in Simmons, 1964 Inbreeding and the mutational and balanced
humans. Genetics 173: 891–900.
loads in natural populations of Drosophila willistoni. Genetics
Fujiwara, T. M., K. Morgan, R. H. Schwartz, R. A. Doherty, S. R.
50: 1299–1311.
Miller et al., 1989 Genealogical analysis of cystic fibrosis fam-
Manatrinon, S., C. Egger-Danner, and R. Baumung, 2009 Estimating
ilies and chromosome 7q RFLP haplotypes in the Hutterite
lethal allele frequencies in complex pedigrees via gene dropping
Brethren. Am. J. Hum. Genet. 44: 327–337.
approach using the example of Brown Swiss cattle. Arch. Anim.
Gerull, B., F. Kirchner, J. X. Chong, J. Tagoe, K. Chandrasekharan
Breed. 52: 230–242.
et al., 2013 Homozygous founder mutation in desmocollin-2
Mange, A. P., 1964 Growth and inbreeding of a human isolate.
(DSC2) causes arrhythmogenic cardiomyopathy in the Hutterite
Hum. Biol. 36: 104–133.
population. Circ. Cardiovasc. Genet. 6: 327–336.
Martin, A. O., 1970 The founder effect in a human isolate: evo-
Gillespie, J. H., 2004 Population Genetics: A Concise Guide. Johns
lutionary implications. Am. J. Phys. Anthropol. 32: 351–367.
Hopkins University Press, Baltimore.
McCune, A. R., R. C. Fuller, A. A. Aquilina, R. M. Dawley, J. M.
Greenberg, R., and J. F. Crow, 1960 A comparison of the effect of
Fadool et al., 2002 A low genomic number of recessive lethals
lethal and detrimental chromosomes from Drosophila popula-
tions. Genetics 45: 1153–1168. in natural populations of bluefin killifish and zebrafish. Science
Haldane, J. B., 1935 The rate of spontaneous mutation of a human 296: 2398–2401.
gene. J. Genet. 31: 317–326. Meagher, S., D. J. Penn, and W. K. Potts, 2000 Male-male com-
Halligan, D. L., and P. D. Keightley, 2003 How many lethal al- petition magnifies inbreeding depression in wild house mice.
leles? Trends Genet. 19: 57–59. Proc. Natl. Acad. Sci. USA 97: 3324–3329.
Hamamy, H., S. E. Antonarakis, L. L. Cavalli-Sforza, S. Temtamy, G. Meyer, W. K., B. Arbeithuber, C. Ober, T. Ebner, I. Tiemann-Boege
Romeo et al., 2011 Consanguineous marriages, pearls and per- et al., 2012 Evaluating the evidence for transmission distortion
ils: Geneva International Consanguinity Workshop Report. in human pedigrees. Genetics 191: 215–232.
Genet. Med. 13: 841–847. Mitchell, K. J., K. I. Pinson, O. G. Kelly, J. Brennan, J. Zupicich et al.,
Hinckley, J. D., D. Abbott, T. L. Burns, M. Heiman, A. D. Shapiro 2001 Functional analysis of secreted and transmembrane pro-
et al., 2013 Quantitative trait locus linkage analysis in a large teins critical to mouse development. Nat. Genet. 28: 241–249.
Amish pedigree identifies novel candidate loci for erythrocyte Morton, N. E., J. F. Crow, and H. J. Muller, 1956 An estimate of
traits. Mol. Genet. Genomic Med. 1: 131–141. the mutational damage in man from data on consanguineous
Hostetler, J. A., 1974 Hutterite Society. Johns Hopkins University marriages. Proc. Natl. Acad. Sci. USA 42: 855–863.
Press, Baltimore. Neel, J. V., W. J. Schull, M. Yamamoto, S. Uchida, T. Yanase et al.,
Howe, K., M. D. Clark, C. F. Torroja, J. Torrance, C. Berthelot et al., 1970 The effects of parental consanguinity and inbreeding in
2013 The zebrafish reference genome sequence and its rela- Hirado, Japan. II. Physical development, tapping rate, blood
tionship to the human genome. Nature 496: 498–503. pressure, intelligence quotient, and school performance. Am.
Huang, L., K. Szymanska, V. L. Jensen, A. R. Janecke, A. M. Innes et al., J. Hum. Genet. 22: 263–286.
2011 TMEM237 is mutated in individuals with a Joubert syndrome Ober, C., T. Hyslop, and W. W. Hauck, 1999 Inbreeding effects on
related disorder and expands the role of the TMEM family at the fertility in humans: evidence for reproductive compensation.
ciliary transition zone. Am. J. Hum. Genet. 89: 713–730. Am. J. Hum. Genet. 64: 225–231.
Hussain, R., and A. H. Bittles, 1998 The prevalence and demo- Overall, A. D., M. Ahmad, and R. A. Nichols, 2002 The effect of
graphic characteristics of consanguineous marriages in Pakistan. reproductive compensation on recessive disorders within con-
J. Biosoc. Sci. 30: 261–275. sanguineous human populations. Heredity 88: 474–479.

Mean Number of Recessive Lethals 1253

R Development Core Team, 2013 R: A Language and Environment edited by J. Crow and J. Neel. Johns Hopkins University Press,
for Statistical Computing. R Foundation for Statistical Comput- Baltimore.
ing, Vienna. Thomas, C. A., 1971 The genetic organization of chromosomes.
Schull, W. J., and J. V. Neel, 1965 The Effects of Inbreeding on Annu. Rev. Genet. 5: 237–256.
Japanese Children. Harper & Row, New York. White, J. K., A. K. Gerdin, N. A. Karp, E. Ryder, M. Buljan et al.,
Schull, W. J., H. Nagano, M. Yamamoto, and I. Komatsu, 2013 Genome-wide generation and systematic phenotyping of
1970 The effects of parental consanguinity and inbreeding in knockout mice reveals new roles for many genes. Cell 154: 452–464.
Hirado, Japan. I. Stillbirths and prereproductive mortality. Am. Wiltshire, K. M., R. A. Hegele, A. M. Innes, and A. K. Brownell,
J. Hum. Genet. 22: 239–262. 2013 Homozygous lamin A/C familial lipodystrophy R482Q
Schurig, V., P. Bowen, F. Harley, and D. Schiff, 1980 The Meckel
mutation in autosomal recessive Emery Dreifuss muscular dys-
syndrome in the Hutterites. Am. J. Med. Genet. 5: 373–381.
trophy. Neuromuscul. Disord. 23: 265–268.
Ségurel, L., M. Wyman, and P. M., 2014 Determinants of muta-
Yampolsky, L. Y., F. A. Kondrashov, and A. S. Kondrashov,
tion rate variation in the human germline. Annu. Rev. Genomics
Hum. Genet. 15: 47–70. 2005 Distribution of the strength of selection against amino acid
Simmons, M. J., and J. F. Crow, 1977 Mutations affecting fitness replacements in human proteins. Hum. Mol. Genet. 14: 3191–3201.
in Drosophila populations. Annu. Rev. Genet. 11: 49–78. Zielenski, J., T. M. Fujiwara, D. Markiewicz, A. J. Paradis, A. I.
Simons, Y. B., M. C. Turchin, J. K. Pritchard, and G. Sella, Anacleto et al., 1993 Identification of the M1101K mutation
2014 The deleterious mutation load is insensitive to recent in the cystic fibrosis transmembrane conductance regulator
population history. Nat. Genet. 46: 220–224. (CFTR) gene and complete detection of cystic fibrosis mutations
Steinberg, A., H. Bleibtreu, T. Kurczynski, A. Martin, and E. Kurczynski, in the Hutterite population. Am. J. Hum. Genet. 52: 609–615.
1967 Genetic studies in an inbred human isolate, pp. 267–290 in
Proceedings of the Third International Congress of Human Genetics, Communicating editor: S. Ramachandran

1254 Z. Gao et al.