Inbreeding Coefficients for X-linked and Autosomal Genes in Consanguineous Marriages in Spanish Populations: The Case of Guipúzcoa (Basque Country)


  • R. Calderón,

    Corresponding author
    1. Departamento de Zoología y Antropología Física, Facultad de Biología, Universidad Complutense, 28040 Madrid, Spain
      *Corresponding author: Prof. Rosario Calderón, Departamento de Zoología y Antropología Física, Facultad de Biología, Universidad Complutense, 28040 Madrid, Spain. Tel: 34 91 394 5111; Fax: 34 91 394 4947. E-mail:
    Search for more papers by this author
  • U. Aresti,

    1. Unidad de Investigación, Sección de Oncología Médica, Hospital de Cruces, 48903 Barakaldo (Bizkaia, Spain)
    Search for more papers by this author
  • B. Ambrosio,

    1. Departamento de Zoología y Antropología Física, Facultad de Biología, Universidad Complutense, 28040 Madrid, Spain
    Search for more papers by this author
  • A. González-Martín

    1. Departamento de Zoología y Antropología Física, Facultad de Biología, Universidad Complutense, 28040 Madrid, Spain
    Search for more papers by this author

*Corresponding author: Prof. Rosario Calderón, Departamento de Zoología y Antropología Física, Facultad de Biología, Universidad Complutense, 28040 Madrid, Spain. Tel: 34 91 394 5111; Fax: 34 91 394 4947. E-mail:


Inbreeding patterns over the past two centuries have been studied more extensively in Spain and Italy than anywhere else in Europe. Consanguinity studies in mainland Spain have shown that populations settled along the Cantabrian cornice share inbreeding patterns that distinguish them from other populations further south. A visual representation of spatial variations of two key inbreeding variables is presented here for the first time via contour maps. This paper also analyzes time trends of mean inbreeding coefficients for X-linked (Fx) and autosomal genes (F) (1862–1995) together with variations in Fx/F ratios in Guipúzcoa, the most autochthonous Spanish Basque province. Because close cousin marriages are a mark of identity of the study population, we evaluated the contribution of uncle-niece/aunt-nephew (M12) and first cousin (M22) marriages to Fx and F values and compared the frequencies of M12 and M22 pedigree subtypes and their corresponding Fx/F ratios to those found in other Spanish populations. The mean Fx and F inbreeding levels in Guipúzcoa for the 134-year period analyzed were 1.51 × 10−3 and 1.04 × 10−3, respectively, and the Fx/F ratio was seen to be very stable over time. Our findings show that major similarities exist for close consanguineous marriage subtypes between Basque and non-Basque Spanish populations, despite significant geographic variability in terms of first cousin pedigrees. The distortion seems to be caused by Guipúzcoa. The Fx/F ratios for first cousins in Spanish populations were higher than expected (1.25), with values ranging from 1.34 to 1.48. The findings of the present study may be useful for advancing knowledge on the effects of the interaction between biology and culture and for exploring associations between mating patterns and the prevalence of certain diseases.


The study of consanguineous marriages using data obtained from Catholic Church dispensation records has shed light on the magnitude and time trends of inbreeding patterns in Spanish populations and territories of very different characteristics and sizes. Inbreeding over the last two centuries has been studied more extensively in Spain and Italy than anywhere else in Europe. While Spain has one of the highest levels of inbreeding in Europe, its rate of consanguineous marriage has declined both later and faster than in other parts of the continent.

The large number of publications which have analyzed consanguinity in Spain to date have yielded a number of interesting findings. First, populations located along the Cantabrian fringe in northern Spain, share certain inbreeding patterns that distinguish them from their more southern counterparts. These patterns are characterized by a comparatively high frequency of mating between close relatives (uncle-niece/aunt-nephew, first cousin and certain types of multiple consanguineous marriages) (Gómez, 1990, 2006; Calderón et al. 1993; Varela et al. 1997, 2001, 2003; Fúster, 2003; Aresti, 2006). The region's relief features, climate, dispersed population (as a settlement model), culture, and history have all played a role in shaping the consanguinity structure of the populations along the Cantabrian cornice. Nevertheless, these similarities in marital behaviour do not seem to extend to genetic variation patterns (Salas et al. 1998; Brion et al. 2004; Calderón et al. 2007). Second, the highest mean inbreeding coefficients (F) in Spain have been observed in the central part of the Cantabrian coastline; in the centre west of the country (around the border separating Castile and Extremadura), and in the centre east (around the border separating Castile and Aragón). Finally, the highest ratios of first cousin (M22) to second cousin (M33) marriages (preferentiality index) have been registered in the Basque Country and the neighbouring region of Cantabria, although high values have also been observed in the inner regions of Galicia and Catalonia, and in southern Andalusia.

Using contour maps, we present for the first time a visual representation of the spatial distribution of mean population inbreeding coefficients (F) and M22/M33 ratios for a set of mainland Spanish populations. The results support the aforementioned points (Figure 1). The consanguinity studies from which we have drawn data cover a wide temporal range (mostly 1900 to 1979), and varying population sizes. The scope of our analysis is obviously limited by data availability but the overall spatial distribution of these two key consanguinity variables in mainland Spain would stand for a good aproximation to a more homogeneous spatial and temporal scenario during most of the 20th century.

Figure 1.

Contour maps showing lines of equal value for the ratios of first cousin (M22) to second cousin (M33) marriages (preferentiality index, M22/M33) and mean population inbreeding coefficient (F) in peninsular Spain. Lines were computer-reconstructed using the Kriging method in Surfer v.8 (Golden Software, Inc). Limit lines for administrative Spanish regions (autonomous communities) are also shown.

To date, little is known about the mean inbreeding coefficient for X-linked genes (Fx) in Spanish or other European populations, or about how it compares to the mean inbreeding coefficient for autosomal genes (F). Research in this field would be of particular interest for furthering knowledge of human population genetics, the relationship between biology and culture, and genetic risk (Barrai et al. 1962; Hajnal, 1963; Hedrick & Parker 1997; Calderón et al. 1995).

The genetic consequences of the structure of pedigrees are important as the pedigree coefficient for X-linked genes (fx) for a female descendent of a biologically related couple ranges from 0 to a positive value, depending on the distribution of the sexes of intermediate ancestors (Hedrick, 2005). Since human consanguinity is influenced by lasting sociocultural rules, which differ from one population to another, these rules may have significant effects on the magnitude of the mean X-linked inbreeding coefficient in the population (Fx).

In this paper we estimated mean X-linked (Fx) and autosomal (F) inbreeding coefficients over a 134-year period in Guipúzcoa, the most autochthonous of the three Spanish Basque provinces. We compared secular trends in Fx vs F coefficients, and analyzed Fx/F ratio variations. As close-kin marriages are characteristic of the Guipúzcoan population, we also evaluated the contribution of uncle-niece/aunt-nephew and first cousin relationships to the above variables. Finally, we compared observed frequencies of pedigree subtypes for both uncle-niece/aunt-nephew and first cousin marriages and their corresponding Fx/F ratios with those registered in other populations in mainland Spain. Our findings might be useful for advancing knowledge on the causes and effects of the interaction between biology and culture in European populations, and for laying the groundwork for exploring in greater depth the promising associations between mating patterns and the prevalence of certain diseases in well-defined populations, such as the Basques.

Material and Methods

Data on consanguineous marriages in the Basque province of Guipúzcoa were compiled from Catholic dispensations records held in the Diocesan Archives of San Sebastián, the capital of Guipúzcoa. Detailed information on the province and its autochthonous population can be found elsewhere (Alfonso-Sánchez et al. 2001, 2005; Aresti 2006).

We analyzed the period from 1862 to 1995 and studied 303 uncle-niece/aunt-nephew marriages and 3178 first cousin marriages. For our general research, we considered the following categories of kinships, named according to Catholic canon law recommendations (codes and pedigree inbreeding coefficients are shown in brackets): uncle-niece/aunt-nephew (M12, f= 1/8), first cousin (M22, f= 1/16), first cousin once removed (M23, f= 1/32), second cousin (M33, f= 1/64), second cousin once removed (M34, f= 1/128), and third cousin (M44: f= 1/256).

Changes in church law during the 20th century had an enormous impact on dispensations requirements for related couples seeking permission to get married. Dispensation, for example, was no longer a requirement for M34 and M44 marriages from 1918 onwards, or for M23 and M33 marriages from 1982 onwards, and nowadays, church dispensation is only required for close-kin marriages (between uncles and nieces, aunts and nephews, and first cousins). Changes in dispensation requests may therefore have led to an underestimation of population inbreeding levels.

A consanguineous pedigree is determined by the sex and number of intermediate ancestors on the two lines. Thus, each consanguinity degree can generate different pedigree subtypes, which can increase in number (shown in parenthesis) the greater the distance between the couple and their common ancestors: M12(4), M22(4), M23(16), M33(16), M34(32), and M44(32). In the present study, pedigrees were coded using the method devised by Barrai et al. (1962).

The formulas used to calculate the pedigree inbreeding coefficients for autosomal (1) and X-linked genes (2) were as follows (see example in Hartl & Clark 1989, p. 240).


is the number of all possible paths through all common ancestors


is m less all paths with 2 or more consecutive males


is the number of individuals in each path


is the number of females in each path


is the inbreeding coefficient of the common ancestor in each path.

The mean inbreeding coefficients F and Fx for the general population were estimated using data available for consanguineous marriages in Guipúzcoa. F was calculated using the formula inline image, where pi is the relative frequency of marriages of i degree of relationship and fi the corresponding pedigree inbreeding coefficient. Fx was calculated using a similar formula to above, but with pi defined as the relative frequency of each subtype of marriage of i degree of relationship and fxi as the corresponding subtype inbreeding coefficient. Four possible pedigrees, for example, can be drawn for a first cousin marriage, each resulting in different pedigree inbreeding coefficients for female offspring fx= 0 (n = 2), fx= 1/8 (n = 1), and fx= 3/16 (n = 1). The autosomal inbreeding coefficient for offspring resulting from a first cousin marriage is f= 1/16. Figure 2 shows the pedigree structure and inbreeding coefficients for uncle-niece/aunt-nephew and first cousin marriages.

Figure 2.

Pedigree subtypes for uncle/niece-aunt/nephew (M12) (upper line) and first cousin (M22) (lower line) consanguineous marriages. Pedigree structures for each category of cousin mating have been classified using the decimal code system devised by Barrai et al. (1962). The pedigree coefficients for X-linked (fx) and autosomal (f) genes are shown.

It should be noted that the Fx/F ratio is a function of the product of the absolute frequencies of the different subtypes of consanguineous marriages (Msti) recorded in a population and the X-linked pedigree inbreeding coefficient (fx); Mi and fi are, respectively, the absolute frequency and the autosomal inbreeding coefficient of a consanguineous marriage with i degree of relationship:


The value of the Fx/F ratio is, therefore, independent of MT (the total number of marriages registered in a population over a period of time).

Statistical analysis of the data contained in the consanguinity database was performed using version 13 of the SPSS statistical software package for Windows.

Construction of Contour Maps

Contour maps were generated from two variables: the ratio of first cousin to second cousin (M22/M33) marriages and the mean autosomal inbreeding coefficient (F) for mainland Spanish populations. The maps were created using the Surfer v.8 geostatistic program (Golden Software, Inc.,Golden, CO, USA) and irregularly located data were interpolated (gridded) using the Kriging method (Delfiner, 1976) which supports breaklines. The database included information on inbreeding data gathered from a selected panel of 46 Spanish populations of different demographic sizes and geographical extensions (covering from large ecclesiastical regions and provinces to small geographical areas and valleys). The database is available from the corresponding author on request.


Table 1 shows the time trends observed in Guipúzcoa between 1862 and 1995 for number of consanguineous marriages, mean inbreeding coefficients (Fx and F), and Fx/F ratios. As can be seen, the highest inbreeding rates were recorded between 1871 and 1920 (Fx= 0.0032–0.0043; F= 0.0023–0.0029). By 1980, consanguineous marriages in Guipúzcoa had become uncommon, representing just 0.01% of all marriages and with an F value of below 0.0001.

Table 1.  Consanguinity, population inbreeding coefficients (Fx and F) and Fx/F ratio in the Basque province of Guipúzcoa (1862–1995).
CohortsaMTbMccM12cM22Inbreeding Coefficients (×1000)
FxFFx/F ratio
  1. aMT total number of marriages

  2. bMc number of consanguineous marriages

  3. cM12 and M22 are, respectively, uncle-niece/aunt-nephew and first cousin consanguineous marriages


In most human populations, the mean X-linked inbreeding coefficient is usually higher than its autosomal equivalent (Christiansen & Feldman 1986; Hartl & Clark 1989). We also found this to be the case in Guipúzcoa, with mean values for the whole period (1862–1995) of Fx= 1.51 × 10−3 and F= 1.04 × 10−3. The differences between the two coefficients were statistically significant (t = 5.52, df = 12, p < 0.01).

Fx and F experienced similar temporal variations, which were, in addition, strikingly synchronic (Figure 3). The sustained similarities between Fx and F over the years might be due to the high proportion of mating between close relatives in both simple and multiple consanguineous marriages. A total of 8388 consanguineous marriages were recorded in Guipúzcoa between 1862 and 1995, and 37.9% of these were between first cousins (1.06% of all marriages). Moreover, maternal pedigree subtypes (fx > 0) were more common than paternal subtypes (fx= 0).

Figure 3.

Secular variations (1862–1995) in the mean inbreeding coefficients F and Fx and the Fx/F ratio in the Basque province of Guipúzcoa.

We found that the relative contribution of first cousin relationship to the overall level of inbreeding in the next generation in Guipúzcoa was high (Fx= 65%, F= 63%); the corresponding figures for uncle-niece/aunt-nephew mating were far lower and very similar to each other (Fx= 12.3%, F= 12%). In a recently published study dealing with the influence of endogamy and consanguinity on genetic disorders in a rural region of Northern Sweden, Bittles and Egerbladh (2005) found similar levels of first cousin marriages (∼2%) to us for a comparable period of time (1860–1899); this type of relationship also contributed considerably to the total inbreeding in the population (F∼ 62%).

The mean Fx/F ratio for all consanguineous marriages recorded in Guipúzcoa during the study period was 1.46 ± 0.02 and this proportion has remained highly stable over the last two centuries (b = 0.0072 n.s.) (see Figure 3). Nevertheless, it is worth observing that the ratio displayed a slight upward tendency between 1950 and 1980 (1951–60: 1.42; 1961–70: 1.64; 1971–80: 1.58), interestingly a period in which large numbers of migrants from other parts of Spain moved to the Basque Country in search of opportunities offered by the growing industrialisation of the region. According to Guipúzcoan census figures from the Basque statistical office (, the population density of the province rose from 187 inhabitants per km2 in 1950 to 319 inhabitants per km2 in 1970. As is well known, changes in population dynamics triggered by favourable economic growth can have a considerable impact on marriage patterns. A study analyzing recent inbreeding in Guipúzcoa showed that the consanguinity rates observed between 1951 and 1995 were influenced by the contribution of immigrants, although differences between the immigrant and the endogamous population for first cousin subtypes were not statistically significant (Alfonso-Sánchez et al. 2001). Consequently, the increase observed in Fx/F ratios in Guipúzcoa from 1950 onwards may be due to an increase in relative frequencies of first cousin unions combined with a sustained excess of maternal pedigrees (fx= 3/16 and/or fx= 1/8).

Hajnal (1963) estimated the Fx/F ratio as a function of the mean difference of age at marriage of husband and wife and other population characteristics (such as whether populations were demographically stable or not or if they were open or closed to migration). For instance, assuming a population with a patrilocal migration pattern (where the probability of staying in the same place is 0.95 for men and 0.75 for women), and a mean spousal age difference of 5 years, the Fx/F ratio for M22 relatives would be 1.56. Thus, the sexes of the intermediate ancestors within the pedigree and the frequencies with which each pedigree type occurs in the population are a result of both the age difference factor and migration patterns in marriage. A similar reasoning is true for M12 unions.

In Guipúzcoa, the estimated mean Fx/F ratios for M12 and M22 relationships were 1.47 and 1.48, respectively and their time trends are shown in Figure 4, which also shows the corresponding proportions for the whole population. As expected, the M22 Fx/F values are very close to those calculated for all degrees of consanguineous marriages registered in the population. The Fx/F ratios for less close relatives were 1.34 for M23 unions, 1.36 for M33 unions, 1.10 for M34 unions, and 1.48 for M44 unions.

Figure 4.

Temporal trend of Fx/F ratios for M12, M22, and all consanguineous marriages in Guipúzcoa over the study period (1862–1995).

Table 2 shows the absolute and relative frequencies of uncle-niece/aunt-nephew (M12) and first cousin (M22) pedigrees in four different populations of Spain; two of the populations are located in the Basque Country (with data from Guipúzcoa and the adjacent province of Alava), one in the Eastern Pyrenees (La Seo de Urgel dioceses, Catalonia) and one in the meseta (Sigüenza-Guadalajara dioceses, Castile). With the exception of the Eastern Pyrenees, the data for the other populations are from long and rather comparable periods of time. As can be seen, there are major similarities for close consanguineous marriage subtypes within Spain (even between Basque and non-Basque populations). In the case of M12 marriages, the cross paternal pedigree subtype (M12–1: 53–65%, fx= 1/8) is by far the most common M12 subtype in Spain. This is probably because the expected age difference between husband and wife is lower that it would be in the other M12 subtypes, and the sex of intermediate ancestor of the female spouse is a woman (there are no intermediate male ancestors in the pedigree, k= 0).The next most common pedigree is M12–3 (26.8–38.2%, fx= 1/4) while the two aunt-nephew subtypes combined range from 3.0% to 14.6%. The frequencies of uncle/niece-aunt/nephew subtypes did not yield statistically significant population differences (χ2= 11.864, df = 9, p = 0.221).

Table 2.  Pedigree subtypes for close consanguineous marriages (uncles-nieces/aunts-nephews, M12 and first cousins, M22) in Spanish populations.
Type of KinshipaCodebfxbfSPAIN
Basque CountryEastern PyreneesCastile
GUIPUZCOA 1862–1995cALAVA 1831–1980dSEO de URGEL 1900–1925eSIGUENZA-GUADALAJARA 1855–1979
  1. aBarrai's et al. (1962); bfx and f are pedigree coefficients for X-linked and autosomal genes, respectively.

  2. cCalderón et al. 1995; ddata provided by A. G-M; eMorales, 1992

Totals  301 97 26 64 
Totals  3160 1170 482 3250 

Maternal first cousin pedigree subtypes (fx > 0) are slightly more common in Spain than paternal subtypes (Calderón et al. 1995; present study) but this difference varies considerably from region to region (χ2= 25.028, df = 9, p = 0.003). The distortion seems to be caused by Guipúzcoa, which has a rather high rate of maternal parallel first cousin subtypes (M22–0: fx= 3/16); the χ2 test yielded a high adjusted residual value. This is not surprising if we consider that Guipúzcoa, and particularly the more rural southern parts of the province, has a higher-than-average rate of inbreeding, presumably for linguistic reasons (Alfonso-Sánchez et al. 2005; Calderón et al. 2006). Even so, this heterogeneity should be considered with caution as it is known that slight differences between populations can easily become statistically significant in large samples (8062 consanguineous marriages in our case). The two Spanish Basque provinces, Guipúzcoa and Alava, showed similar first cousin subtype patterns (χ2= 3.98, df = 3, p = 0.26). Table 3 shows estimated Fx/F ratios for the same biological relationships and populations as above. The Fx/F ratio was slightly more variable between populations for uncle-niece/aunt-nephew (1.34–1.55) than for first cousin unions (1.34–1.48). The expected Fx/F ratios based on uncle-niece/aunt-nephew and first cousin subtypes would vary in both cases from 0 to 3, with a maximum being reached when subtypes with the highest fx pedigree inbreeding values (M12–0, fx= 3/8 and M22–0, fx= 3/16) yielded frequencies of 100%. In contrast, if the observed frequencies of these pedigree subtypes for each of the kinship categories were identical, the estimated Fx/F ratios would be 1.5 (M12) and 1.25 (M22). As can be seen in Table 3, the Fx/F ratios for first cousins are higher in all cases than 1.25. A review of variations in Fx/F values among a number of worldwide human populations can be found elsewhere (Calderón et al. 1995).

Table 3.  Variations in Fx/F ratios among different Spanish populations.
Ecclesiastical Regions (Dioceses)Geographical AreaPeriodinline image
  1. aM12 and M22 are, respectively, uncle-niece/aunt-nephew and first cousin relatives.

GuipúzcoaBasque Country1831–19801.471.48
AlavaBasque Country1862–19951.341.40
Seo de UrgelEastern Pyrenees1900–19251.441.41
Expected values  1.501.25


Inbreeding among Basques has not reached exceptional levels in the last two centuries. Nevertheless, the consanguinity structure in the Basque area is different to that of other Spanish regions in that it is characterized by a strong, sustained preference for mating between close relatives (e.g. M22/M33 > 1) and relatively high rates of maternal cousin marriage. Also noteworthy is the fact that inbreeding peaked during the first decade of the 20th century but has been in sharp decline ever since. This pattern—which has also been observed to a greater or larger extent in other parts of Western Europe—has been attributed to the early industrialisation and urbanisation of a large part of the Basque Country at the end of the nineteenth century (Calderón et al. 1993; Alfonso-Sánchez et al. 2001, 2005; Calderón et al. 2006).

It has been widely demonstrated that religious and sociocultural factors have a major impact on human mating systems. Consanguineous marriages in the Middle East, India, South-East Asia, Japan, and sub-Saharan Africa, for example, are still common and constitute a considerable proportion of all marriages. In many of these regions, with mainly Muslim or Arab populations, consanguineous marriages constitute a third of all marriages, and there is a strong preference for first-cousin unions (particularly paternal parallel subtypes, M22–3, fx= 0) (Khoury & Massad 1992; Modell & Darr 2002; Bittles, 2002; Meyer, 2005; Hamamy et al. 2007). There is, however, a growing tendency towards other close cousin marriage subtypes (Hamamy et al. 2005). In contrast, marriage patterns in Europe have traditionally been less influenced by sociocultural factors than have those in Eastern societies. This relatively low preference for certain consanguineous marriage subtypes could be broadly extrapolated to groups characterized by a deeply rooted cultural background in which family ties have remained relatively strong over generations. A good example is the population of Guipúzcoa, a province located in the heart of the Basque area which, has historically housed the highest rate of autochthonous Basques and speakers of the Basque language (Calderón et al. 1998a; Alfonso-Sánchez et al. 2005).

Morton (1955) and Barrai et al. (1962) showed that the sex of intermediate and immediate ancestors of a related couple can have a considerable influence on certain genealogical models in human populations. The moderately high frequency of maternal consanguineous marriage subtypes (even for distant degrees of consanguinity) in the Guipúzcoan population might be due to a rather matriarchal tendency shown within Basque society, whereby the mothers of the couple would have taken a greater interest in the marriage and this, in turn, would have strengthened family ties and improved the social and economic relationships within the family nucleus (Dueso, 1990).

This scenario may have influenced Fx/F ratios. As Table 3 shows, all the Fx/F ratios for uncle-niece/aunt-nephew and first cousin subtypes were greater than 1 and the magnitude of estimates varied spatially within a limited range. Based on these and other observations taken from the literature, the average population Fx-values in humans might be more intimately related to sociocultural factors than mean F-values at autosomal loci. Differential effects on each of these inbreeding coefficients, however, would be less noticeable in Europeans than in Asians. This would seem to indicate certain homogeneity of attitudes in terms of preference for or avoidance of certain subtypes of cousin pedigrees, although this would not be the case for the main consanguinity types.

Current inbreeding levels are low in Guipúzcoa, as they are in other European populations. This does not, however, mean that we will witness an immediate decrease in the rate of deleterious recessive genes or in the frequencies of genetic diseases because of the complex evolutionary processes required to achieve equilibrium of gene frequencies. Moreover, given that the tendency towards this equilibrium is very slow, it is unlikely that allele frequencies in human populations are balanced with respect to their historic or current inbreeding levels (Christiansen & Feldman, 1986, p.166).

The incidence of recessive autosomal disorders in the general population is regulated by the homozygote value aa=q2+Fq(1 −q), where q is the disease allele frequency and F the mean population inbreeding coefficient. Consequently, a change in the level of inbreeding (caused, for example, by a change in mating systems or a decrease in population size) would immediately alter the frequency of a given disease in a particular cohort or generation. For example, for low F values (which are typically observed in European societies today), the frequency of a recessive allele in an inbred population is defined as inline image, where μ is the mutation rate for the recessive allele, and t is the selection coefficient. The formula shows that the frequency of recessive alleles, and hence the incidence of disease q2 will increase in the future. The effect of using the average inbreeding coefficient (F) rather than the equilibrium inbreeding coefficient (Fe) to calculate the equilibrium frequency (qe) of a recessive disease in a given population, is negligible for Europeans but considerable for highly inbred populations (such as those in India), as was established by Hedrick (1986). As Fx > F then from the previous formula it follows that qeX < qeA if μ/t is the same in both cases (where qeX and qeA are the equilibrium frequencies for X-linked and autosomal loci, respectively). The expected lower equilibrium for X-linked deleterious alleles when it is compared with that for autosomes is a prediction that may be of important significance for genetic disease frequencies.

Viscoso and Charlesworth (2006) state that the hemizygosity in males may cause differences, between the X chromosome and the autosomes, in patterns of evolution including rates of gene divergence and pattern of gene expression. They discuss evidence suggesting that both positive and negative selection act more efficiently at X-linked loci. However, the precise nature of those processes remains to be elucidated in future research.

Inbreeding levels increased considerably in Europe during the second half of the 19th century but began to decline sharply from the early 20th century onwards, to return to below the initial level (Sutter & Tabah, 1948; Sutter & Goux, 1962; Moroni et al. 1972a, 1972b; Pettener 1981; Jorde & Pitkänen 1991; Calderón et al. 1993; Cavalli-Sforza et al. 2004; Blanco-Villegas et al. 2004; Bittles & Egerbladh, 2005). This time trend has led to contrasting results in terms of qe values. As there have only passed few generations (approximately three), evidence of these changes is probably scarce, but the overall tendency may be an increase in the incidence of recessive disorders. Current research into the impact of inbreeding in humans is also investigating reproductive compensation in consanguineous couples to gain a better understanding of disease incidence, allele frequency, and consanguinity in general (Overall et al. 2002). A review of the cost of human consanguinity and its relevance to clinical genetics is provided by Bittles et al. (1991) and Bittles (2001)

At the population level, studies designed to evaluate the biological effects of inbreeding have mainly centred on the offspring of first cousin marriages, because they are the most common of all marriages between close relatives, and are also an important contributor to the frequency of rare recessive Mendelian disorders. First cousin marriages are, in this sense, very similar to certain types of multiple consanguineous marriages (involving two or more types of kinship) which are common in certain populations, such as the Basques and Castilians. Multiple consanguineous marriages often have high pedigree coefficients, but insufficient information on such unions is available due to underreporting (Calderón et al. 1998b).

The frequency (K) of a recessive genetic disorder in the offspring of first cousins compared to that in the offspring of non-related couples is determined by inline image (Dahlberg, 1930), where c is the observed frequency of first cousin marriages in the population and F, the mean inbreeding coefficient. The frequency of X-linked disease is obtained using a similar formula: Kx=c(q2+ (1 −q) q f22X)/(q2+ (1 −q)q FX) where f22X is the pedigree inbreeding coefficient for first cousin subtypes and FX is the mean population inbreeding coefficient for X-linked genes. For empirical values of c and q, Kxc (q+Fx) (see Calderón et al. 1995). Interestingly, recent consanguinity studies have suggested that population substructures need to be taken into account to accurately estimate disease incidence (Overall et al. 2003).

The main difficulty in most human populations lies in estimating disease allele frequency (q) rather than disease frequency (K). To address this problem, research that has attempted to calculate the cost of consanguinity in humans in terms of the incidence of recessive disorders has been applied to large, extensive populations in which inbreeding patterns have been analyzed for long periods of time (Romeo et al. 1983, 1985). This and similar strategies have shown that studying the population as a whole, rather than just individuals or specific families, is essential for the analysis of health-related events in human populations.

Recent scientific findings have provided interesting epidemiological data on Mendelian, or complex, multifactorial, disorders detected in certain Basque families or in random samples of the autochthonous population. Examples of specific Mendelian recessive diseases studied in Basques include cystic fibrosis (CFTR gene) (the ΔF508 mutation is more common in Basques than in other European populations; Morral et al. 1994; Casals et al. 1997); limb-girdle muscular dystrophy type 2A (LGMD2A gene) (the 22362AG→TCATCT mutation seems to be relatively common in Basque patients from Guipúzcoa and the age of the mutation has been estimated at 50 generations; Urtasun et al. 1998; Cobo et al. 2004); HFE-gene related hereditary hemochromatosis (two mutations are significant in Basques: the C282Y mutation, which has been found in low frequency in autochthonous patients with family origins in Guipúzcoa and the French Basque Country, and the H63D mutation, which is more common in Basques than in most European populations; de Juan et al. 2001; Bauduer et al. 2005). Other research works dealing with the expansion of a trinucleotide repeat [CGG]n located within the FMR1 X-linked gene, as the main cause of fragile X syndrome, Peñagarikano et al. (2004) conclude that the prevalence of potentially unstable alleles among Basques is similar to that of other Caucasian populations. Finally, hereditary haemophilia (factor XI deficiency), is also relatively common in the Basque area (the Cys38Arg mutation has yielded a frequency of 0.005 in French Basque patients; Bauduer et al. 2002) and its presence seems to be due to genetic drift (founder effects) (Zivelin et al. 2002). In Spanish families other cases of X-linked recessive disorders have been observed: Hunter disease (Gort et al. 1998), Retinitis pigmentosa (García-Hoyos et al. 2006) and Wiskott-Aldrich syndrome (Andreu et al. 2003) among others.

No studies to date have analyzed the association between breeding patterns and the prevalence and incidence of genetic diseases in Spain as a whole, the Basque Country, or any other major Spanish regions. The findings of such studies would be of considerable research interest. Given the high quality of current and historic data available for inbreeding patterns and levels in the Spanish Basque area and the weight of diseases in certain, well-defined populations, it would be important to evaluate the impact of consanguinity on rare or more common recessive disorders in autochthonous Basques and non-Basques (migrant groups) living in the same geographic region. As mentioned before, Guipúzcoa, together with the other two Basque provinces, Vizcaya and Alava, received large numbers of migrants from other Spanish regions several decades ago (1950–1970). In addition, given that the substructure within populations has an important effect on the inbreeding coefficient of any generation, and that the genetic analysis of a population depends upon knowledge of mating systems, it would also be very interesting to unveil the existence or nonexistence of geographical disease patterns in the Basque area which might mimic the existence of the well-known genetic substructure of the Basques (Aguirre et al. 1991; Calderón et al. 1998a). Such research should be enriched with information concerning the demographic history of the population, which has influenced gene frequencies of different types of diseases through founder effects, effective population sizes, population growth, and times of population expansion. Such histories, which are associated with key evolutionary parameters such as those mentioned above are usually complex and many related aspects remain unknown to us (Li, 1976; Lange & Gladstien, 1980; Sankaranarayanan, 1998; Yokoyama, 1983). The clear identification of relationships between all these sources of information will mark a considerable advance in knowledge in a field of enormous anthropological and medical interest, with important implications for health policy and genetic epidemiology.


R. Calderón thanks the ecclesiastical authorities responsible for the Archives of San Sebastian Dioceses for providing permission and all facilities to study consanguinity in Guipúzcoa. U. Aresti received a fellowship of the Basque Administration to carry out his PhD on this subject at the University of the Basque Country. This study was a component of a large project funded by the Basque Government (grant no, GVPI/1998-2001/18) to RC.