Tuesday, November 26, 2013

DNA shared averages with multiple relationships.

The below chart shows the averages of autosomal DNA shares between relatives (modified from the one at http://www.isogg.org/wiki/Autosomal_DNA_statistics)

  • 50% parent/child or full sibling
  • 37.5% "3/4 siblings"
  • 25% grandparent/grandchild, full aunt or uncle/niece or nephew, half-siblings, double first cousins 
  • 12.5% great-grandparent/great-grandchild, first cousins, great-uncles or great-aunts/great-nephew or great-niece, half-aunts/uncles, half-nephews/nieces
  • 6.25% first cousins once removed, double second cousins
  • 3.125% second cousins, first cousins twice removed
  • 1.563% second cousins once removed, double third cousins
  • 0.781% third cousins, second cousins twice removed
  • 0.391% third cousins once removed, double fourth cousins
  • 0.195% fourth cousins
  • 0.0977% fourth cousins once removed double fifth cousins
  • 0.0488% fifth cousins

We all know the averages aren't exact, and get less exact the further the relationship, and by the time you get to 5th cousins it varies widely, but lets look at some interesting (to me anyways) possibilities for other relationships.

"3/4 siblings" (credit to CeCe Moore for coining that phrase, at least thats the first person I have heard use it meaning half siblings with the not-shared parent a full sibling to the others not-shared parent) could appear the same as a full sibling, as they would share 25% +12.5%  which totals 37.5%, but would be the only relationship other than full siblings that should show significant fully indentical segments.

Double second cousins would share 3.125% X 2 = 6.25% which is the same as first cousins once removed.

Double third cousins would share .781% X 2 = 1.562% which is the same as 2nd cousins once removed.

Double fourth cousins .195% X 2 = 0.39% which is the same as third cousins once removed.

2nd cousins who are also double 3rd cousins 3.125% + 1.563% = 4.688%

One would expect the shared DNA to be in multiple segments.  23andMe routinely assigns people to estimated 4th cousinship because they share 2 segments but often this is a segment on each parent. (my personal experience, yours may vary)