March 16, 2007

Muslim and Christian Lebanese or Hasty Conclusions in Human Population Genetics

On my post on Genetic Palimpsests, I argued that genetic interpretations about historical events are often suspect when done with very old markers:
Often, this historical reasoning can be shaky. For example, Spencer Wells has made tall claims about the Phoenicians, the Sea Peoples, and the Carthaginians in a National Geographic article which were based on the analysis of haplogroup J and E distribution in the Levant and North Africa.
The reason for this is the following. Suppose a marker is very old, much older than the historical phenomenon under study. In the case of the above-mentioned documentary, the ages of the markers used were about 7 times greater than either the appearance of the Sea Peoples or the Carthaginians. Thus, if we observe similarity between two populations based on such old markers, we cannot conclude that they have a common recent historical origin. This is why Spencer Wells' inspired study on The Phoenicians was wrong when it claimed:
The tests could confirm that men of Tyre-Christians and Muslims alike--are related to the ancient traders. Wells and Zalloua also took samples in other parts of the Phoenician world, where results may reveal the same lineage in areas of former colonies like Sardinia and Malta.

...

During the bloody civil war of the 1970s and 1980s, some groups used the name Phoenician as an ideological weapon. Certain Maronites, the dominant Christian sect in Lebanon, claimed a direct ancestry from the Phoenicians, implying that they held a more legitimate historical claim on Lebanon than later immigrants from the Arabian Peninsula. This inflamed many Muslims. The term Phoenician had turned into a code word for Christian rather than Muslim.

...

Could genetics show that modern Lebanese, both Christians and Muslims, share the same Phoenician heritage? That's one question this project, funded by the National Geographic Society, hopes to resolve.

...

That result delights Zalloua; it supports his belief that both Muslim and Christian Lebanese populations share an ancient genetic heritage.

"Maybe now we can finally put some of our internal struggles to rest," he says.

It is clear from this article that the two possible outcomes of the study were not treated equally: the idea that Muslims and Christians are alike "delights", while the idea that they are not is demonized as contributing to sectarian violence and civil war in Lebanon.

Science should be about the facts, not about wishful thinking. The science of population genetics does have the potential to be used for political reasons, but that is no excuse for drawing unsubstantiated conclusions based on what is politically useful.

The results of the National Geographic study were never published in a peer reviewed journal, so they did not undergo the normal process of scientific scrutiny. However, it is clear why the results were not sufficient: the markers used were ancient. One cannot conclude a recent origin of two populations sharing a common ancient marker, because differences can be revealed when other, more recent, markers are used.

Indeed, in a different study (pdf) which studied more derived markers, the difference between Christian and Muslim Lebanese was made apparent:
The PC plot suggested the presence of four main groups (Fig. 3a): 1) North Africa, Figure 2 Genealogical relationships of the selected UEPs. Nomenclature as suggested by YCC (2002). 2) Near East/Arabs (including Muslim Lebanese and Ashkenazi Jews), 3) Central-East Mediterranean grouping, including Christian Lebanese and 4)West Mediterranean.
What is the reason for these conclusions? The most striking difference between the Christian and Muslim Lebanese is within haplogroup J, i.e., the haplogroup supposed to reflect their common heritage. Muslims have 56.4% of J, while Christians have 44.2%, but this is distributed 30.8%/25.6% among haplogroups J*(xJ2) and J2 in Muslims and 9.3%/34.9% among Christians. It is the high frequency of J*(x2) which indicates the substantial Arab ancestry among the Muslims compared to the Christians. So, indeed Christians appear to descend more from the pre-Arab populations of Lebanon, and presumably the Phoenicians, compared to the Muslims who are more similar to other Arabs.

The literature is full of similar hasty conclusions. The Cohen Modal Haplotype debacle is another case in point. In this case, a simple 6-locus STR haplotype in a YAP- background was taken to be indicative of Aaronic biblical ancestry, a conclusion which did not withstand further scrutiny. At present, 10 years after the Y-DNA Aaron proposition made its appearance, no evidence in support of this theory has been presented; customers of genetic testing companies dabbling in "Jewish ancestry" are expectedly at a loss.

Similarly, haplogroup R1a1 has been proposed as indicative of Viking, Slavic, Kurgan, Ukranian Paleolithic, Indo-Aryan, etc. expansions although very little is known about its phylogeny since the Upper Paleolithic.

Of course R1a1 may have hidden phylogenetic structure that could be linked to the various proposed population expansions, and Aaron's Y chromosome may have been inherited by some bearers of the CMH. But the case needs to be made.

No comments: