Ancestral Haplotypes

I just answered a researcher's question about how to “... identify the apparent haplotype of the MRCA”. 

In case you don't know what the ancestral haplotype is - it is the marker values that the ancestor would have had in order for his descendants to have the values they have  - which must be mostly the same as the ancestor's values - but possibly with one or more random mutations mixed in.

This researcher had already picked up on my use of "apparent" - which I use because the haplotype is derived from the available data and is not necessarily correct - though it is probably either correct or very close.  She also used a very common abbreviation "MRCA" which stands for Most Recent Common Ancestor - which would be the closest ancestor to the men who is a direct ancestor of all of the men in the group. 

Here are the 3 basic approaches to determining an "Apparent Ancestral Haplotype":

  1. The ideal way to do an analysis for "Apparent MRCA Haplotype" is to have 3 sets of results - with each man descended from a different son of the MRCA.  In this instance, the “2 out of 3” rule works – if at least 2 of the men have the same value – that means that the MRCA almost certainly had that value in order to pass it through two of his sons.  Any different value is the mutation.  Some researchers call this "Triangulation".
  1. When there are no known paper trail connections, we use the “majority rules” approach.  My semi-automated Results Tool “declares” an apparent ancestral value only when at 75% or more of the values are the same at a marker – and provides a “?” when there isn’t the clearly dominant majority value for a marker.  (this approach can sometimes lead to flawed assumptions when most men tested are in the same family – as their results can dominate – whether they actually represent the apparent ancestral value or not)
  1. When there are some known paper trail connections, it may be possible to use that insight to do a more appropriate job of picking out the apparent ancestral value.  (This would become a "hybrid" between approaches 1 & 2)  For example, if there are 10 men being evaluated and 6 descend from the same ancestor, first create an apparent ancestral haplotype for the ancestor of the 6 - and then use his apparent haplotype  along with the other 4 haplotypes to define an overall Apparent Ancestral Haplotype for the full 10 man group

No matter which of these 3 approaches you are using, please keep in mind that you should continually reasses the "Apparent Ancestral Haplotype" as you obtain more data.