The chance method to testing hypotheses that I described within the “Science the Heck Out of Your DNA” sequence depends on the underlying values being correct. The calculations are primarily based on this graph from AncestryDNA’s Matching White Paper.
Every curve exhibits how possible the attainable relationships are for a given quantity of shared DNA. For instance, when two individuals share 200 cM, the connection is most probably in Group E or Group F, with a lot decrease probabilities that it might be in Teams D or G.
Whereas the shapes of the distributions (curves) are in all probability correct, the graph could not totally symbolize the very extremes of every relationship group. That’s, there actually exist “outliers” for every class which can be so uncommon that the chance within the graph seems to be zero when it’s actually, say, 0.002.
If you’re eyeballing relationships, your mind makes allowances for attainable outliers (in all probability too many allowances!), however a computerized calculator, just like the one described within the “Science the Heck” sequence, can’t. A price of zero is at all times zero, and when multiplied by different values nonetheless offers zero. That signifies that in very uncommon circumstances, information from an outlier might rule out an speculation that’s truly true.
To enhance the calculator, I’m accumulating information on confirmed outliers. These are matches for whom the connection might be confirmed by different DNA proof however who share centimorgan quantities outdoors the identified ranges. For instance, if two first cousins match each other as anticipated and likewise match their youngsters as anticipated, however their youngsters don’t match each other as 2nd cousins, that’s information that might profit the group.
For those who suppose you will have an outlier instance, please run the shared DNA quantity by the Shared cM Software with Chances. If the identified relationship just isn’t listed as an possibility, please report the outlier utilizing this on-line type.