A new study by Gilbert et al. (2017) examines the population structure among Irish Travellers, a small community within Ireland (29,000–40,000 individuals) that have a nomadic history, much like the Bedouins and the Roma (Romani) people. Due to an increased urbanization process the Irish Travellers became increasingly urban although the maintained their language (Shelta) and social cohesion mainly through consanguineous marriages.
The Donoghue twins sit inside their caravan at the Ballinasloe horse fair in County Galway, Ireland. Source
To study the population structure the authors carry out PCA on haplotypes inferred by fineStructure and find that the Travellers appear distinct (why don’t get to see the PCA on the actual data, for reasons that may become clear later). However, seeing these plots in other papers, the most obvious question is not why the the Travellers are distinct but rather whether matching Irish individuals were included in the analysis. We discussed how PCA plots can be misleading in Marshall et al. (2016). This question may be difficult to answer with the information provided in the paper.
The authors also do unsupervised-admixture (K=4-6), where they find the Traveller to have an ancestral component that doesn’t exist in other populations.
Repeating the analysis with Roma populations (K=2-4) the mysterious component appear to be shared with Settled Irish and Roma.
What does it it mean? Well that’s the beauty part about admixture analyses. It means absolutely nothing, or in other words, it means whatever it is you want it to mean. Since the two admixture analyses are not comparable neither in the number of components nor in the populations being compared we are left asking what exactly the authors wanted to think, or what we are not supposed to see if all the populations would have been studied together along all K’s? The authors could have also read recent papers criticizing the approach of showing multiple Admixture analyses with various K’s and asking the reader to pick the best one and device an alternative strategy. Either way, the population structure remains unclear.
The final analysis is aimed to find the best fitting parameters for the demographic inference. It is delighting to read that a bottleneck has both been assumed (based on what exactly?) and confirmed (with a model devised to identify bottlenecks), much like in the Zidan et al. (2014) paper on the Druze, which we criticized (Marshall et al. 2016).
Overall, this paper deals with an interesting question, but the analyses are confusing and inconclusive. Hopefully, future studies would continue the effort to shed light on the history of the Irish Travellers. Until then, it is interesting to read the National Geographic’s story about the Irish Travellers (Source).
Genomic insights into the population structure and history of the Irish Travellers
Edmund Gilbert, Shai Carmi, Sean Ennis, James F. Wilson, and Gianpiero L. Cavalleri
The Irish Travellers are a population with a history of nomadism; consanguineous unions are common and they are socially isolated from the surrounding, ‘settled’ Irish people. Low-resolution genetic analysis suggests a common Irish origin between the settled and the Traveller populations. What is not known, however, is the extent of population structure within the Irish Travellers, the time of divergence from the general Irish population, or the extent of autozygosity. Using a sample of 50 Irish Travellers,
143 European Roma, 2232 settled Irish, 2039 British and 6255 European or world-wide individuals, we demonstrate evidence for population substructure within the Irish Traveller population, and estimate a time of divergence before the Great Famine of 1845–1852. We quantify the high levels of autozygosity, which are comparable to levels previously described in Orcadian 1st/2nd cousin offspring, and finally show the Irish Traveller population has no particular genetic links to the European Roma. The levels of autozygosity and distinct Irish origins have implications for disease mapping within Ireland, while the population structure and divergence inform on social history.
Gilbert, E., S. Carmi, S. Ennis, J. F. Wilson, and G. L. Cavalleri. 2017. Genomic insights into the population structure and history of the Irish Travellers. Scientific Reports. 7.
Marshall, S., R. Das, M. Pirooznia, and E. Elhaik. 2016. Reconstructing Druze population history. Scientific Reports. 6:35837.
Zidan, J., D. Ben-Avraham, S. Carmi, T. Maray, E. Friedman, and G. Atzmon. 2014. Genotyping of geographically diverse Druze trios reveals substructure and a recent bottleneck. European Journal of Human Genetics. 23:1093-1099.