Thursday, July 8, 2021

Inbreeding Declined In West And Central Eurasia During The Holocene Era

A new paper published in the journal Cell shows that inbreeding has declined over the last 12,000 years, as demonstrated with 411 ancient DNA samples from West Eurasia (defined as Europe) and Central Eurasia (defined as Southwest Asia, the Caucasus and Central Asia).

Extreme consanguineous matings did occur among agriculturalists but were rare, while inbreeding was more common in hunter-gatherers and declined more gradually as farming societies grow more complex. Modern Europe, which has far more people, has even less inbreeding.

The regressions and images above go from the present on left, to the past, on the right. F(ROH) (for runs of homozygosity) is a measure of inbreeding with a higher number indicating more inbreeding in the lineage of the individual whose DNA is examined.

Essentially, the larger the size of the communities involved, as improving food production technology makes possible, the less inbreeding is found.

The detailed numerical data is summarized in the chart below. The key data points (with median F(ROH)) are:

* Hunter-gathers (N=40) 0.0633
* Simple agriculturalists (N=102) 0.0286
* Early complex agriculture (N=230) 0.0250
* Advanced complex agriculture (N=160) 0.0160
* Modern Central Eurasia (N=309) 0.0156
* Modern Europe (N=139) 0.0039

Central Eurasia is less inbred than Europe among hunter-gathers, simple agriculturalists, and advanced complex agricultural societies, but is modestly more inbred in the era of early complex agriculture, although the statistical significance of the differences in these areas is modest.

In the modern era, Central Eurasia with an inbreeding coefficient of 0.0156 (which is more than in Central Eurasia during the advanced complex agriculture phase) is much more inbred than modern Europe which has an inbreeding coefficient of 0.0039. The combined modern inbreeding coefficient is 0.0066.

Prevalence Of Moderate Inbreeding With Some Cousin Marriage

100% of hunter-gatherers, 95% of farmers prior to the advanced complex agriculture phase, 77% of farmers in an era of advanced complex agriculture, and 14% of modern Europeans are as inbred as the median modern Central Eurasian (0.0117). This is what one would expect to arise from endogamy within a community with some cousin marriage (including remote cousins) that is not predominant.

Prevalence Of Cousin Marriage

93% of hunter-gatherers, 5-15% of ancient farmers, and 2% of modern Europeans are as inbred as 23% of modern Central Eurasian are (0.0391), a level typical for first cousin marriages. According to the paper:
Samples from modern groups like the Balochi, the Bedouin, or the Sindhi from Pakistan have the highest proportions of individuals with FROH > 0.0391 (50%, 41.3%, and 33.3% respectively).
Prevalence Of Highly Inbred Mating

10% of hunter-gathers, 1.9% of simple agriculturalists and modern Central Eurasians, 0.4% of farmers in early complex agriculture, and 0% of farmers in advanced complex agriculture and modern Europeans has an inbreeding coefficient of 0.0932

This is a level of inbreeding in excess of otherwise unrelated double cousin marriages, and similar to that of marriages between an uncle or aunt and that person's niece or nephew, between half-siblings, between a grandparent and a grandchild, or offspring of closer matings. 

Even remote cousin over many generations in a small community, however, can elevate the inbreeding coefficient of mere cousin marriages or of a double cousin marriage, above this threshold. The genetic evidence indicates that this sort of genetic drift was involved in the early hunter-gatherer communities, while cousin and/or closer marriage was a major factor in other genomes studied.

The high inbreeding coefficients in modern "Central Asians" are largely a product of Islamic law's acceptance of cousin marriage and of traditional cultures in some (but not all) Islamic societies, that favor high rates of cousin marriage of 10% to more than 50%, as shown on the map below. 

Cousin marriage is uncommon, however, despite being at least nominally legal under Islamic law, in countries that are predominantly Muslim where Muslims predominantly follow the school of Sunni Islamic jurisprudence that is predominant in Southeast Asia (i.e. the Shafi'i school of Sunni Islamic law).

Some concluding comments of the paper bear repeating in their entirety so as not to lose the nuance of these important caveats to its findings:
Three points further deserve mention regarding mating patterns in human societies. 
One is the seeming contrast between the high levels of drift-driven autozygosity (panmictic inbreeding) we report for ancient hunter-gatherer societies and ethnographic studies showing low levels of inbreeding among modern-day hunter-gatherers. For instance, a comparison of inbreeding patterns in a worldwide sample of contemporary hunter-gatherers with Amazonian horticulturalists reported lower inbreeding in hunter-gatherer groups. Hill and colleagues also report low levels of relatedness within modern-day hunter-gatherer bands. However, the mentioned ethnographic findings rely on genealogies and report the prevalence of inbreeding by consanguinity, not inbreeding by drift. In fact, we also find consanguinity to be rare among early Holocene Eurasian hunter-gatherers relative to agriculturalists, consistent with widespread exogamy in modern-day hunter-gatherers. This raises the possibility that reciprocal exogamy and consanguinity avoidance traditions may have been predominant among human foragers since prehistory (but possibly not in archaic hominins).

Second, our results lend support, albeit with limited data, to the hypothesis that extreme consanguinity may have become more common with farming. This result parallels higher within-group marriages among modern-day horticulturalists than foragers. It is also consistent with singular reports on ancient agriculturalist genomes, such as evidence for consanguinity identified in an early Neolithic farmer from Iran, a first-degree incest case from Neolithic Ireland, as well as a recent report on close-kin unions in the central Andes after 1000 CE. 

In our analysis, among the seven individuals with the highest level of inbreeding (with FROH > 0.125), all four hunter-gatherers appear autozygous by drift, while all three agriculturalists appear autozygous by consanguinity. This appears unlikely to happen by chance (Fisher’s exact test, two-sided p = 0.029). These results are consistent with the view that consanguineous traditions could have thrived in class-based agricultural societies with private property more readily than in more egalitarian hunter-gatherer groups.

Finally, we report higher consanguinity in Central versus West Eurasia in contemporary societies, in parallel with earlier work. This is consistent with widespread first- or second-cousin marriage practices in agricultural societies in Middle Eastern and North African countries and in South Asia, including Muslim and Jewish groups, as documented by ethnographic or genomic studies. We note that cousin marriages were also common among royal dynasties and upper classes of Europe until the 20th century, and many prominent European scientists of that period are known to have married their first cousins, including Charles Darwin and Albert Einstein. These traditions are thought to have arisen through various social factors, including the inheritance of property in class societies.

Interestingly, we do not observe the relatively high rates of consanguineous marriage observed in modern-day Central Eurasia in any of the past societies we studied, in Antiquity or earlier. We naturally prefer to remain cautious, especially given the limited sample size of our advanced complex agriculturalist samples from West and Central Eurasia (n = 9 and n = 30, respectively). Nevertheless, it appears possible that present-day cultural patterns may have emerged relatively late in time.


Onur Dincer said...


Cousin marriage is uncommon, however, despite being at least nominally legal under Islamic law, in countries that are predominantly Muslim where Muslims predominantly follow the school of Sunni Islamic jurisprudence that is predominant in Southeast Asia (i.e. the Shafi'i school of Sunni Islamic law).

Most Kurds and South Indian Muslims and many Arabs are Shafi'i and cousin marriage is common among Shafi'i Kurds, South Indians and Arabs as well. So I think we should search the roots of the non-prevalence of cousin marriage among SE Asian and East African Muslims at other factors.

andrew said...

"Most Kurds and South Indian Muslims and many Arabs are Shafi'i"

In the case of Kurds, this is somewhat incomplete as they are one of the most religiously diverse predominantly Islamic ethnicities: "Most Kurds are Sunni Muslims who adhere to the Shafiʽi school, while a significant minority adhere to the Hanafi school. Moreover, many Shafi'i Kurds adhere to either one of the two Sufi orders Naqshbandi and Qadiriyya. Beside Sunni Islam, Alevism and Shia Islam also have millions of Kurdish followers." Marriage practices among Kurds also show great variation: "Barth finds in his study of southern Kurdistan that in tribal villages 57% of all marriages were cousin marriages (48% bint 'amm marriages) while in a nontribal village made up of recent immigrant families only 17% were cousin marriages (13% bint 'amm)." So, without further breaking down marriage rates by religion among Kurds, it is rather hard to tell.

The Shafi'i Arabs are a minority Sunni Muslim population in most of the places in the Levant where they are present, sometimes in a fairly complex religious environment with Shi'ites, Druze, Christian Arabs, and Jews also present, so it is not a very straightforward matter to match cousin marriage rates among Arab Muslims in these places either.

I don't necessarily mean to say that Shafi'i doctrine itself discourages cousin marriage. I have no idea one way or the other on that point. But, my intuition is that this school of Islamic law is largely tracing some shares cultural and/or socioeconomic traits that often manifest is low cousin marriage rates. Determining whether that is the case, or whether the alignment between the school of Islamic law and local cultural marriage practices is almost purely a coincidence that isn't consistent across the school is something that the data in the places you mention isn't unequivocal and clear enough to establish. You could be right, but I'm skeptical of that possibility.

andrew said...

I do concede your point in Yemen where Arabian Peninsula ethnic practices appear to prevail: "Raphael Patai reports that in central Arabia no relaxation of a man's right to the father's brother's daughter (FBD, or paternal female cousin) seems to have taken place in the past hundred years before his 1962 work. Here the girl is not forced to marry her paternal male cousin but she cannot marry another unless he gives consent. Among the Jews of Yemen this rule is also followed albeit not as rigidly. In northern Arabia the custom is very strong and any outsider wishing to marry a woman must first come to the paternal male cousin, ask his permission, and pay him what he wants, and a man who marries off his daughter without the consent of the paternal male cousin may be killed by family members. The right of the paternal male cousin is such that a *shaykh* may not be able to prevail against it. Among the Bedouin it can happen that a paternal male cousin can lodge a complaint after the marriage has taken place, compelling the father to reimburse the bride price or have the marriage annulled. If the paternal male cousin cannot marry his paternal female cousin immediately due to financial or other considerations, the paternal male cousin can also "reserve" her by making a public and formal statement of his intentions to marry her at a future date. A more distant relative acquires priority to marry a girl over her paternal male cousin by reserving her soon after her birth."

Onur Dincer said...


I live in a country in which Kurds are the biggest minority (Turkey) and know plenty of Shafi'i Kurds from real life. I can assure you that cousin marriage is quite common among them, especially in the older generations, who tend to be more traditional. Also, there is nothing discouraging cousin marriage in Shafi'ism (or in any other major Islamic school of law). If a certain Muslim population has low to no cousin marriage, it is almost surely because of preservation of pre-Islamic marriage customs. Some more counter examples to your Shafi'i theory: Dagestani Muslims are Shafi'i and commonly practice cousin marriage, Balkan Muslims, Circassians, Central Asian and Caucasian Turkic peoples, Tatars, Bashkirs and Hui are by and large Hanafi and in all of them cousin marriage is uncommon. Can we argue based on these examples that Hanafism discourages cousin marriage? Of course not as we know from the Hanafi doctrine and from many other mostly Hanafi peoples among whom cousin marriage is common. Similarly, the Shafi'i doctrine does not discourage cousin marriage and the examples I have provided from mostly Shafi'i peoples who commonly practice cousin marriage are in agreement with this.

Onur Dincer said...

Another point, no mainstream Islamic school of law discourages cousin marriage and they simply cannot do so. Because Koran does not prohibit it and, more importantly, many early Muslims married one or more cousins, including Muhammad himself, according to their official Islamic biographies. Some of the imams of the mainstream Islamic schools of law made some negative remarks about cousin marriage due to its ill effects, but because of the accounts of early Muslims practicing cousin marriage they did not insist on that position and tolerated it.