June 06, 2009

YHRD updated germline mutation rates

YHRD has updated with new mutation rates for Y-STR loci. This is an invaluable resource, as it pools mutation rates from all published studies to produce an overall rate based on a large sample. 

As I have argued elsewhere, uncertainty about the mutation rate is one of the major sources of uncertainty about age estimates of Y-chromosome common ancestors. Hopefully more researchers will start using these rates (as e.g., Di Gaetano et al. did) rather than those picked from single papers of the literature, especially "evolutionary" ones.

4 comments:

Gioiello said...

When it has been permitted, in the previous posting on “Sicilian Y-chromosomes: Greek and North African influences”, to someone Abate Vella to say: “Latin language was never spoken in Sicily by the population at large. It was the language of the Roman rulers and of a small part of the cultural elite. Even with regard to the Sicilian language it is not at all sure its placement among the Latin languages”, I think that any confrontation becomes difficult.
This is a short sentence in Sicilian:

E’ megghju futtiri ca cumannari o è megghju cumannari ca futtiri?

Translated in Italian:

E’ meglio fottere che comandare o è meglio comandare che fottere?

(Is it better to fuck than to rule or is it better to rule than to fuck?)

Someone can translate it in Greek, in Arab, in Maltese (if Mr Vella is more a Maltese than a Sicilian) and everybody will be able to judge.

Ponto said...

My surname is Scottish. In Scotland there were five ethnicities each with their own language: The Strathclyde Welsh, the Scotti, the Picts, the Sassanachs and the Norse. The language of the Sassanachs prevails today. In Scotland you don't find people waffling on about language change or testing for haplogroups to find the true Scot and the immigrant Scot. A large proportion of the Scottish population is of recent Irish immigration and Roman Catholic, yet no-one in Scotland is having pogroms over it. Other people, Europeans and Americans could learn from the Scots.

I use the average mutation rate of 0.002 per generation irregardless of the marker involved. Five markers or 67 markers, it does not matter. It is an average that works well for me. For instance, I am J1, the only J1 with DYS578=7. According to the mutation rates, the rate is 0.00008. That rate is quite ridiculous and would mean I am totally unrelated to any other J1 just on that marker. You have to use common sense. What is probably true is that most common clades for European haplogroups today date back to about 4 kya, and are not really usable in estimating ultimate origin or migration paths. J1 for instance is found in diverse lands in Africa, Europe and Asia, older than any ethnic groups claiming it today, and to boldly state it originated on the Arabian Plate? as it does in Wikipedia is nonsense, but that is basically what you get in genetic studies, a lot of nonsense.

The Greeks who went to Sicily could have been haplogroup I or R1b, you don't really know. We are just guessing using the haplogroups and clades found in Greece and Sicily today and making assumptions on the past.

McG said...

I will agree, now, that Zhivs rates are too slow. I found that Chandlers germ-line rates worked well in estimating TMRCA for a scottish clan of known historical origin. The biggest problem I had was distinguishing between close relatives and not overcounting mutations. I did that by excluding some of the faster mutating dys loci. The second most important issue was not counting multi-step mutations as multiple steps. Since about 5% of all mutations are multi-step, this can significantly increase measured Variance. This factor can produce errors by itself of up to 100%! Finally, I hope they plan to expand the list of dys loci?

Ponto said...

There are accepted ways of working out genetic differences for markers with more than one copy like DYS389. For example, 13 and 30 compared with 12 and 29. This is a GD of 1: [13-12] + [(30-13) - (29-12)]. If in doubt leave out multi copy markers and use the rest. As I said I use an average of 0.002/generation for all markers which effectively flattens out the fast and slow mutation rates. In my 67 markers, where I differ from my genetic cousin, no paper work but same surname, are in fast and slow markers, each marker a GD of 1. Using individual marker mutation rates would give a wide range of possibilities to our common ancestor. Using the average gives a reasonable number of generations and years to our common ancestor.

Use whatever system works for you.