This is a copy of the first BanyanDNA newsletter, delivered 9 December, 2023. You can sign up for the newsletter here.
Welcome to the inaugural newsletter for BanyanDNA, the next generation of DNA analysis!
BanyanDNA is unlike any other tool for genetic genealogy. Not only can it help you identify an unknown parent, grandparent, or great grandparent, it can alert you to places in your tree where shared DNA does not support the documented relationships. Best of all, it is completely customized to your family, whether you have pedigree collapse, double cousins, or (in a future release) endogamy.
In the example below, Ann is an adoptee with an unknown birth mother. However, her closest DNA match, Jane, is the child of first cousins who married. No other tool can analyze a double-cousin scenario like this one, but BanyanDNA can!
With an intuitive interface, BanyanDNA makes it easy to build a tree for Ann’s matches and to see where Ann best fits in the tree. Each possible location is called a hypothesis.
With just four DNA matches and a few seconds of calculations, BanyanDNA can tell us that Ann is probably the granddaughter of Olive Hemmings (Hypothesis 1). Now we know where to focus our efforts to find Ann’s mother!
The official launch of BanyanDNA will be on 29 February 2024 at RootsTech in Salt Lake City. You can register to attend online or in person here: https://www.familysearch.org/en/rootstech/
We are planning a soft launch for our biggest fans (you!) a few weeks beforehand. Stay tuned to our monthly email newsletter for details, and please spread the word.
Looks very interesting. I hope you aren’t kidding about “intuitive.”
We’re trying to make it as intuitive as possible!
How does this differ from the What Are The Odds (WATO) tool found at DNA Painter (www.dnapainter.com)?
Many genealogists will continue to be served well by WATO. BanyanDNA can do two things WATO can’t: it can handle double relationships, and it can analyze the fit of the DNA matches across the entire tree (not just between the target and matches).
Sounds interesting if it delivers
This looks so exciting 🙂
Can it find my 4th grandmother? I have found th main segment she’s on but can’t tell if it is her or her husband. She had no daughters. I can find no one among her known descendants that is no one older than me who had tested on this line that helps.
Ultimately, autosomal DNA is limited by biology. No tool can change that, I’m afraid.
I can’t wait to see how this works and how it can be used in my twisted, double cousin-ed, endogamous tree!
How is this different from the What are the Odds tool?
The concept is similar, but BanyanDNA can do a few things that WATO can’t, including double relationships.
Very Very interesting.
I am eager to see it…
Can’t wait
I am looking forward to working with this new tool.
I am so excited for this tool to launch! Just giddy!
is this an online tool only or one that can work on own PC? It looks interesting. Also what are likely costs, and other details?
It’s an online app that works through your internet browser. We haven’t worked out the final pricing model, but users will be able to create one project for free.
Hopefully this program will be helpful with my adopted daughter’s situation. We know her bio parents, but 3 of her 4 grandparents are mysteries. I’m working with DNA Painter’s WATO as well, and I suspect several incidents of siblings raising nieces/nephews as their children. Fortunately I have a decent supply of DNA matches.
Will it have the possibility to add haplogroups and take them into consideration for the calculations?
That’s definitely something we’re thinking about!
Very exciting…especially the upcoming expansion for endogamy!
Unable to attend RootsTech in person (first time unable to since it started) – what is going to be best way to learn your tool?
Are there any future online webinars planned?
We’ll have a quickstart video, extensive documentation online, plus a Facebook group to help people learn the tool. I’ll also be offering classes after RootsTech; gotta get through that first!
Interesting as I have some NPE “cold cases” which would need systematic analysis. However their difficulty lies therein that the NPEs occurred in 19th century, so mostly 4-6 generations ago and match quantity and even more connected genealogical knowledge is very scarce.
I’m trying to understand the concept of BanyanDNA from the homepage:
– building a tree of known (and suspected) relations of persons of interest
– entering as much match data (cM) between persons of interest as available
– not sure if the data provider field does influence the BanyanDNA algorithm?
– Calculate probabilities of hypotheses
In my case for some NPE I was lucky to build one hypothesis from the genealogy (I reconstructed) of often only one usable DNA match. I’m however struggling to find or formulate another hypothesis. This is mainly trough limited genealogical knowledge, meaning not the full ancestry of the DNA match is known.
So my question would be if using only one hypothesis will also work and give a statistical evaluation of probability?
And what if no relationships hypothesis can be build at all? Can BanyanDNA use the cM to an “unknown unconnected match” (to multiple persons in the connected tree) to give a likelyhood of relations and their probability?
A tip for the homepage: clearly formulating what BanyanDNA can achieve for users in the opening banner text might help the project. Currently there is an explanation of the name Banyan (I would not do that in the most important homepage space) and that family trees can be similar. Hence the problem. But there is no short explanation what is the solution to the problem and how it is achieved with BanyanDNA.
Interesting, as I have some “non-paternal event” (NPE) “cold cases” which would require systematic analysis. However, the challenge lies in the fact that the NPEs occurred in the 19th century, which means they happened 4-6 generations ago. Match quantity and even more importantly, connected genealogical knowledge, is quite scarce.
I am trying to understand the concept of BanyanDNA from the homepage:
– building a tree of known (and suspected) relations of persons of interest
– entering as much match data (centiMorgans) between persons of interest as available
– not sure if the data provider field does influence the BanyanDNA algorithm?
– Calculating probabilities of hypotheses
In my case for some NPE, I was fortunate to build one hypothesis from the genealogy (I reconstructed) of often only one usable DNA match. However, I am struggling to find or formulate another hypothesis, mainly due to limited genealogical knowledge, meaning not the full ancestry of the DNA match is known.
So my question would be if using only one hypothesis will also work and give a statistical evaluation of probability?
And what if no relationships hypothesis can be built at all? Can BanyanDNA use the centiMorgan (cM) data to an “unknown unconnected match” (to multiple persons in the connected tree) to give a likelihood of relationships and their probability?
A suggestion for the homepage: clearly stating what BanyanDNA can achieve for users in the opening banner text might help the project. Currently, there is an explanation of the name Banyan (I would not do that in the most important homepage space) and that family trees can be similar. Hence the problem. However, there is no short explanation of what the solution to the problem is and how it is achieved with BanyanDNA.
Hi Chris, thanks for the feedback on the website. It’s still a work in progress, so your suggestions are appreciated. We’ll post detailed documentation soon which should answer most of your questions.
I am excited and anticipating the launch of your program. I think it will be helpful for my complicated genealogy.
The whole team is looking forward to seeing how it helps you!
Can someone upload a GEDcom file, or do they need to build the tree from scratch? Being able to download a GEDcom of the tree that was made in the program would be helpful too. I’m working on entering my tree.
BanyanDNA does not accept gedcoms at the moment. We have the capability, but depicting complex trees in a 2D format is proving to be a challenge.
I would need to be able to import a Gedcom file with 100,000 plus linked individuals going back to the 1300’s.
Projects are currently limited to 500 individuals, although you could import specific sublineages of your gedcom as long as the number of people stayed below that limit.
UPDATE: Gedcom imports are now available to our premium users as an opt-in beta feature. The positioning algorithm still needs work, though, so some nodes may be placed on the screen in odd places.
Dr. Larkin,
Forgive me if you have answered this question already. My impression is that BanyanDNA runs the simulations to compute the probability curves for each relationship scenario in the tree. If so, does that mean that BanyanDNA can handle more distant relationships than WATO which (again – my impression) relies on previously generated curves that did not address more distant relationships? Is there a ‘cut off’ relationship or amount of shared cM where BanyanDNA predictions are no longer generated or lose accuracy? I’m thinking of the 40cM average ‘rule of thumb’ for WATO.
Thank you!
Good question! We haven’t tested BanyanDNA for more distant relationships, but I don’t expect it to extend our range much beyond what WATO can do. That’s because once you get down below about 50 cM, there are too many possible relationships for statistical confidence. We do plan to support Y-DNA and mtDNA, eventually, as well as longest segment. Those additions might improve things somewhat.