Coincidentally, while testing the best methods for using multiple siblings’ kits, a person I’m helping has a close match to a sibling pair. This is a practical application of helping people with advanced genetic genealogy methods.

I was recently finishing up an article about using multiple siblings’ kits to approximate a parent’s DNA. I had already done all of the analysis. And then I was contacted by a family that wanted some genetic genealogy help. A man wanted to know who his father was and I was glad to assist him.

He had tested at 23andMe along with his mother and grandfather. I noticed that the next highest match was labeled “mother’s side.” But I knew that couldn’t be right because she only shared 6 centiMorgans (cM) with his mother. There’s a good chance that this is a false match, but even if it wasn’t, people are often related in multiple ways, especially when a match is around the 10th great-grandparent level. These were strong matches on his father’s side.

And then I noticed something very fortunate. This match also had a full-brother tested on 23andMe. What good timing! This was a perfect opportunity for helping people when siblings have tested.

For the two strong matches of the man I was helping, the brother had a nickname and a last initial. The sister had a first initial and a last name that was not the same as the brother’s, so obviously a married name. I asked the person I was helping to send a message to these matches, but before that could happen I found everything we needed to know. Through some sleuthing on the internet, I was able to find the sister’s full name and then her maiden name fairly quickly. This maiden name had the same initial as her brother. That turned up a strong possibility for his name as well.

Using advanced research techniques

A year ago I developed a method for using the kits of multiple children as a very good predictor of how much DNA their applicable parent would share with a cousin. More recently I’ve done extensive testing on that method. Here are the articles that can help you do the same:

Since this match was at 23andMe, I was able to get the distinct segments that the two children share with the man I was helping. It was made extremely easy with a tool that Jonny Perl, which I discuss in detail in Part 3 of the series above. I’ve used this tool hundreds of times lately for testing and a few more times to solve genetic genealogy cases recently. The tool gave a total amount of 602.4 cM.

This number should not be used on its own. It would be fine for three or more siblings, but not for two. If these matches were half-siblings, it would have been very easy to then multiply 602.4 cM by (100 – the % shared between the siblings). But, since they’re full-siblings and I don’t have access to their kits, I would have to multiply by 100/75, which is the inverse of the percentage of a parent’s DNA two children will have, on average. The former method only has the slightest improvement over the 100/75 method, but is a whole lot more work for full-siblings.

Recall that the cM from distinct segments between the two siblings and the man I was helping added up to 602.4 cM. Multiplying that by 100/75 results in 803.2 cM.

Taking two matches that are about 400 cM and instead getting 803 cM for their father is a huge improvement. In this case, the result isn’t much different from multiplying the average cM times two, but in many other cases that method isn’t nearly as accurate. Part 2 and Part 3 of the series on these methods show just how big the differences are.

I put the value 803 cM into the most accurate relationship predictor. The result was that the most likely option was that the father of the siblings was a full 1st cousin to the man I was helping, with a probability of 32.3%. However, I could tell that a 1st cousin of the father would likely be born between about 1900 and 1930. Nearly all of these men were already deceased. I knew that that wouldn’t preclude them from being the father of the man I was helping, but the ages made them seem like unlikely candidates compared to the alternative.

So I concentrated on the next highest probability, at 29%, which was that the man I was helping is a half- or great-nephew to the father of the siblings.

Building trees

Most of the two and a half hours I spent on this case went towards tracking down obituaries and entering the information into a family tree I created. I could have tried to wing it with notes or tried to remember some of the information, but that would’ve been folly. You never know when you need to come back to something. You could end up following an incorrect lead, in which case you’ll need to be able to backtrack and consult a tree. Starting over from scratch would not be fun.

I found another match at 23andMe that was shared by the man I was helping and both of the siblings. The last name of this match strongly indicated that the shared connection was on the siblings’ father’s side and was probably a 2nd cousin once removed or 3rd cousin. I had already built a tree for both the father and the mother at that point, but it also looked like the siblings’ mother had no siblings of her own. It turned out that I already had the tree built farther back than I needed to.

Wrapping it up

It turned out that the father I was looking for should be a nephew of the father of the two sibling matches. I had already built this part of the tree first across and then down. There were three men who were nephews to the father of the sibling matches, which meant that their son would be a great-nephew to the father of the two siblings. I soon sent an email to the person I was helping. I gave the following three options:

  • The name of a man who lives 700 miles away
  • The name of a 50 year-old man living 500 miles away in a very different direction
  • The name and age of another man of a particular age living in the same state

A couple of hours passed by, which was long enough for me to begin to question my findings. But, no sooner had I begun to look over it again, when I got a reply.

It turns out that the mother of the person I was helping had given a first name already. I didn’t know this at the time, but the first name was the same as the third option I had given.

It wasn’t long before the man I was helping sent a couple of messages to the man who is likely his father. This man denied that he was in a particular place during a particular time. That’s a very common reaction. But it’s difficult for a person who’s always wondered about their parents to hear that once they finally find out who they are. He doesn’t want anything from his father. But of course he was going to try talking to him. And I don’t think he’s done trying either.

I helped him using methods that other people haven’t had access to, but hopefully now that I’ve written a three-part series on these methods, other genetic genealogists will have the same advantages that I have. A 27-year mystery was solved! But what’s left is the hard part—getting cooperation from the father. It’s likely that he knows he’s the father. His DNA matches show that he’s the father. But ideally he would get a DNA test to remove any doubt.

Interestingly, this wasn’t the only case in which I’ve been able to take advantage of these methods. Every case I’ve worked on lately has had two close relatives tested at the same generational level either on the side of the person I’m helping or with their closest matches. I hope that genetic genealogists will help themselves by taking advantage of the most advanced and accurate methods for a scenario that’s actually pretty common, when other close relatives are tested. Just because I’m the one who developed these methods doesn’t mean that I should be the only genetic genealogist who can benefit from them. Any genetic genealogists who’s good at what they do will employ these methods.

Cover photo by Pok Rie. If you had access to the most accurate, free, relationship predictor, would you use it? Feel free to ask a question or leave a comment. And make sure to check out these ranges of shared DNA percentages or shared centiMorgans, which are the only published values that match peer-reviewed standard deviations. Or, try a tool that lets you find the amount of an ancestor’s DNA you cover when combining multiple kits. I also have some older articles that are only on Medium.