# Intrepretation of DNA detectable relationships

While quite a number of questions have been asked and answered regarding DNA and relationships, I'm not certain that this question has been asked.

I have a 5th cousin that has been identified by documentary trail as well as DNA (at 25 cm according to Ancestry.com). That cousin has 5 siblings, and all 6 of these persons are confirmed to be of the same parentage (confirmed by assertion of parents and DNA testing).

According to ISOGG Cousins chart there is (nominally) a 70 percent probability that a 5th cousin is not detectable. That leads me to conclude that using DNA alone, the likelihood is that I might only detect about 2 of the 6 cousins as being a DNA match to me.

Is that a reasonable conclusion?

## 2 Answers

Yes, even if all six are 5th cousins, some will share more DNA than you with others and some may not be detected.

Here is a helpful graphic on coefficients of relatedness indicating approximately how much DNA you will share with different types of relatives on average. You can also look at the table under the "Human relationships" section of this "Coefficient of relationship" Wikipedia article

As you can see from this coefficient of relatedness table fourth cousins share about 0.2% of their DNA. 5th cousins will share even less. Then add in the fact that this is only an average (due to an element of random chance involved in how genetics works, some of your 4th cousins will share less than 0.2% of their DNA with you and some will share more.

With such small numbers (0.2% is a very small amount of DNA!) the distinction between 4th cousins, 5th cousins, etc are a lot less reliable statistically. For example, second degree relatives share about 25% of their DNA but some people that are second degree relatives might share a couple percentages more or less than this. So when you are trying to distinguish between 4th and 5th cousins and dealing with partial percentages its a lot less exact because a certain amount of shared DNA could be indicative of several different degrees of relatedness and genealogical records would have to be used to give you the exact degree of relatedness.

• Could you please make the labels for your linked content descriptive instead of saying "this answer" "this table" and so on? It's more friendly to readers who may be reliant on screen readers or for whom scrolling over the link to reveal the destination takes a lot of effort. If the destination website is in the label, readers know right away whether the content is new to them or a site they've already visited. Jan 8, 2020 at 8:07
• Ok updated the answer, thanks for the tip! Jan 8, 2020 at 17:38

If they were all independent, then yes, I think your analysis would be correct.

However, your cousins are not all "independent trials" (as a statistician would put it).

• I think I understand your qualification, what I'm trying to reconcile is why among the same sibling group some have a detectable relationship to me, why others do not. The "take away" (for me) is that the absence of a DNA match does not preclude a cousin relationship.
– BobE
Jan 6, 2020 at 18:31
• That is correct. Note that Ancestry.com has an arbitrary cut-off point (I think it's 6cm, but I could be wrong) below which they just say "no relation" Jan 6, 2020 at 19:11