Forming and Using Genetic Networks with Genetic Affairs


– [Blaine] Hi, my name is Blaine Bettinger and I want to talk to you very briefly about forming and using genetic networks, using a new tool called Genetic Affairs. This is going to be just a
brief introduction to this tool, but hopefully it will get
you enough information to get started and to start
exploring this tool on your own. So the goal of this tool is to take matches and shared
matching and form clusters, clusters that might look
something like this. This is an example of clusters we’re going to work with in this example. This is the output of the tool. And so, what we’re going to do is learn how to create this output and then learn how to use this output to understand what these clusters are and how they might help us
with our genealogical research. So we wanna take this
output that looks like this and get to a point where it
looks something like this, where we’ve identified the possible line of each of these clusters and can potentially use
that to tell us about the relationship to individuals
within those clusters, both to ourselves and to each other. So the way we create this cluster is using a tool called Genetic Affairs. It’s at geneticaffairs.com. And you can either
register for a new account or if you’re already registered, you can click on members login. I also wanna note that, when
you’re on this initial page at geneticaffairs.com, if you click on the FAQ there at the top that will also give you a manual which has step-by-step instructions and pictures, and as I understand it is being updated as new information is
provided and available, so that could be very
helpful for you as well. So, once you have an account, and an account is free at Genetic Affairs, there are some paid options
that we’ll talk about including for the clustering. But once you’re there,
you can then add a profile to the website that enables
you to create these clusters. Because obviously, you’re
going to have to give it the input it needs to be able
to create these clusters. And the input it needs
is access to your account at either 23andMe, AncestryDNA and Family Tree DNA at the current time. So, what we can do is for
example, add another website here. And what we’re gonna do
is we are going to add an AncestryDNA profile to this account. So we click on add AncestryDNA account and that will take us to a login page. So what we’re gonna do is
we’re going to provide it now with our login information
and our password. And of course, we always
wanna be very cautious about granting our login information. In fact, many of the
company terms of service require us to keep that confidential. So we obviously don’t go share
it, these passwords publicly. One thing you could do for example, that people have suggested, is to create a free account at Ancestry that you share your DNA profile to, and then you could provide
that new free account, that login information, here. And so, what that means is, at any time you can revoke the DNA sharing to that free account you’ve created and thus, you can always be in
control of your DNA account. That, of course will
only work at AncestryDNA because they have that DNA sharing option. Once you’ve provided login
and password information, you’ll get a pop-up
that says, are you sure, we’re going to check the credentials. You click yes. And when that happens, then you’ll see that it’s being checked. And once that’s done, then you’ll see you will have information for your accounts attached under websites here. So you can see, I have
accounts at Genetic Affairs for ancestry DNA, for 23andMe and I also have some Family Tree DNA. I’ve had to add each
of those individually. So, it will recover all of
your AncestryDNA profiles that are associated with your account, people that are sharing
with you, for example. Same for 23andMe. At Family Tree DNA you
have to add of course, each of your individual profiles one by one to provide those. Once you have a number of… Well, once you have even one profile, you can create now, these cluster charts. So what you would wanna do
is you will wanna to click under where it says profiles
here, either here for ancestry, here for 23andMe, or
here for Family Tree DNA. You want to click that,
and then it will take you to a page that lists
your various profiles. Maybe it’s just one, maybe it’s multiples. In order to create this cluster chart, you can see where it
says auto cluster there, click on that rounding arrow. Click on that and then that
will take you to a page where you can begin to
create this auto clustering. Now you can see here, there
are several different ways to select thresholds. Now, you could just use the default and click perform analysis there, or you can change the matching thresholds. Here, I’ve provided the various thresholds for each of the high
and low thresholds here. So, you can select those, if
you don’t know what to pick then just stick with the defaults. You can come back and
change those defaults to get more matches, fewer matches, matches of different sizes and so on. Now, I will warn you that every new run does take credits, which we’ll talk about. But, you can come back and
adjust those thresholds. Once you’ve selected the
thresholds you’re going to use, here, I’ve selected 900 for
a high and nine for a low. Meaning, it’s going to
look for shared matching with matches in your match list that fall between 900 and
and nine centimorgans. So for example, a sibling
wouldn’t be included in this because they share too much DNA with you, above that 900 centimorgan threshold. So, you’ll click on
perform analysis there. And then you’ll see, it will ask you, are you sure you want to
perform this analysis? It’s going to cost you 25 credits. So you can say, well no, cancel or you can click yes, perform
this auto cluster analysis. Now on the issue of credits, you can go back to the member
page or the initial page and you can click the prices icon there. And what you’ll find is it will tell you that
there are various fees associated with Genetic Affairs
if you choose to use them. So for example, if you wanna
get a monthly update on someone meaning, Genetic Affairs
will scan your match list and we’ll tell you what
new matches you have at a certain threshold that you set, that costs 18 credits per profile. A weekly update is nine
credits per profile meaning, you’ll get a
notification once a week for any profiles you select this for. And if you wanna get an update
every day for a profile, meaning, it will look and
scan your matches every day to see what new matches you
have, that’s three credits. And an auto cluster
analysis costs 25 credits. So, credits are the equivalent
of a dollar in US dollars. And so, a $5 would be 500 credit. So in other words, an
auto cluster analysis is about a quarter, 25 cents. And currently, when you sign up for a new account at Genetic Affairs, you’ll get about 200 credits for free, which will allow you to do
several auto cluster analyses. Okay. So, once you see that you
have chosen and auto cluster, you’ll get a notification
that it’s successful and it will tell you that it takes about 10 minutes to perform, in which time it will send the results to your email. So you’re not gonna be looking at the Genetic Affairs
website for your cluster, you’re actually going to be
finding it in your email. You will receive an
email with several files, a couple of them are .CSV
or spreadsheet files, and one of the files ends with .HTML. So when you get the email, it
will look something like this. Here you can say, I’ve gotten
the email at my Gmail account and it says, for profile
xyz, whatever the name is, an auto cluster analysis
has been performed, access it through the attached HTML file. So you can see here, I get three files. One with a list of matches, one with a list of shared matching and then the final one,
which is the .HTML. Now don’t open this from your email. Instead, save it to your
desktop and open it from there. You wanna open it in your browser, okay, this is going to be
happening in your browser. So you open it in Chrome, Firefox, Edge, Internet Explorer,
whatever your browser is. For most of us, 60% of
us it will be Chrome. Once it opens, you can see
it is initially unclustered, it’s just all of these matches. When you scroll down this page
here, what will happen is, it will then automatically cluster. So here for example, now
this is a new cluster. So it goes from what we see
here to what we see here. Now we have seven clusters, orange, green, red, purple and so on. And so what this is, is each cluster is a group of shared matches. So you can see here
that there are reports, 21 people in cluster two. So what that means is,
there will be 21 people that are compared, going
this way and this way. And in the middle of
course, this dark line, is where a person matches themselves. And here, you can see that,
let’s pick this person. This person matches almost everyone in this shared match cluster, except for a few missing people out here. Whereas let’s say, there are
some people like this person which matches absolutely everyone in this shared match cluster. But clearly, most of the people match most of the
individuals in the cluster, so this is a very tight cluster. So, if you scroll down this
page, what you’ll see is, it will tell you that you use thresholds of 900 centimorgans for a high and nine centimorgans for a low. And it identified a total of 92 matches that it used in that clustering. So, now I can say well, alright, that’s a good number, I
can use this analysis. Or if I want, I can go back
and change the thresholds and try it over again. Now, when I scroll down even further, this is the real bulk of the information that’s provided in the auto cluster. And that is, it will tell me
the names of the individuals that are in each of these clusters. So for example, this is showing me information about cluster five,
cluster six, cluster seven. And it will tell me
each of the individuals that are within that cluster. So if we look at this, for example, we can see in cluster five, it reports a couple of individuals there and that corresponds to
that brown cluster five. So the real value of the clustering is the chart at the bottom that tells you who is inside each of these clusters. Now, once we have these
clusters, they look great. They’re fantastic visualizations, they’re really interesting but, how do we actually use these? What are they telling us, and how can we use what it tells us in our genealogical research? Well, it’s important to note first of all, that clusters are clues only. And I cannot repeat that enough. This is not proof of anything. It is certainly DNA evidence, but it is important to
note that it is clues only. It requires a lot of follow-up work to ensure that what we
hypothesize about these clusters is the most likely reason
for the clustering. And the hypothesis for a cluster is this: A group of shared matches are likely or possibly related
through a common ancestor. That’s the basis for genetic networks, that’s the basis for shared matching, that’s the basis for in common with, that people within a
cluster, within a group, are likely related
through a common ancestor. And so, finding the common ancestor for some of the people in a cluster, then suggest the common
ancestor for all of them. Again, this is just a
clue, it’s a hypothesis, it is not proof on its own. So if I can look at a
cluster and I can find the common ancestor for some
of the people in that cluster, that then suggest that the
other people in that cluster are related to me through
the same ancestry. So, let’s look at an example of this. So what we wanna do then, is we want to review the trees of a
cluster, one of the time to look for common ancestry. So we take cluster one, look at the trees of cluster one and try to find a pattern. Are there a common
ancestry in that cluster, among the trees of that cluster, that suggest what the relationship is? And then the common
ancestry that we identify, if we’re able to, is the clue
or the arrow for that cluster. So, let’s look at this. We’re gonna look at
cluster number two first. This is an example of a cluster diagram from Genetic Affairs,
and there to the right is the tree of the individual that took the DNA test
to create this cluster. What we’re gonna do is,
we’re going to look at cluster number two
here, the green cluster. So, what do we do? Well, the first thing we do
is, we scroll down the page to find the table for cluster number two. So what do we do here? This is an Ancestry account,
so what I’m gonna do is, there are 11 people in this cluster, so the very first thing I’m gonna do is, I’m going to open all 11 profile
matches for this cluster. So I’m gonna open each one of
those, all at the same time. I’m gonna have 11 new tabs that open up. And as you can see, some of
these people have a tree. Now what this will show you is, here, for example, all of these people that have numbers here have a tree, and that’s the size of their tree. But note, this does not report people that have an unlinked tree. So I could have more trees
here than what I see, this won’t report those unlinked trees. So open all of these
profiles, every single one. And what we’re gonna do is, we’re simply going to review
the trees of these individuals, if there are any. Now, you’re gonna have some small clusters that have no trees, and that’s
going to be a lot harder, ’cause you’re gonna have to find ways to ask them for trees or
to create trees and so on. But if there are trees, and we have many in this particular cluster. What I do is, I open
all of those, I look for trees that have common
surnames, common locations, any clues, any patterns, any relationships among the people here, that’s
what I’m gonna be looking for. Because once I find that, that’s gonna be the clue or the arrow that points me towards the origin or the reason for the cluster. Why does this cluster exist? Because of relationship
through a common ancestor. Here, when I open up all of these trees, I find in several of them but not all of them, which is common. I find that many of them have the Colwell family in their tree. So therefore, I can tentatively identify this cluster
as a Colwell cluster. What does that mean? It means that this test
taker is likely related to the individuals here in green, through the Colwell family. And so what that means is, the people that have
Colwell in their tree, well, I could have
already hypothesized this. The value to this really, is that the people that didn’t
have Colwell in their tree but are clustering within
that green cluster, that therefore suggest they’re related to the Colwell family. And I could contact them and I can say, do you have the Colwell
family in your tree and so on. And now I’m actually giving
them useful information instead of simply saying, we’re related, do you wanna compare information? And now, I have a focus
for our relationship? So, as I said before,
this tells me therefore that the people in this cluster that don’t have Colwell in their tree are likely related through
this Colwell family. It is not proof, but it is a clue, a very very powerful, very strong clue about our relationship. So what do we do, we
just rinse and repeat, we do this process now
for all the other clusters and we try to figure
out where those clusters line into our tree. A lot of them, you’re not
going to be able to decipher where they fit in with your tree. Potentially because the
matches might be distant, they might not have trees
associated with them, they might not have good trees, they might not have accurate trees. And so, what you can do is, you can do the legwork
to try to figure it out. You can build trees for these individuals. You can use documentary evidence to figure out how they might
be related and so on. So, it often requires
a lot of this legwork. But the goal is to get to
something that looks like this. Now you can see here, I
have six of these clusters identified and tied into the tree. And I wanna also note, it’s not uncommon to have multiple clusters
for a particular line. So for example here, you can see I have a couple of them that
are the Spicer family, a couple that are the Colwell family. And so, what that may represent is, people that are just
clustering differently, or they might represent different portions of the Spicer family. So maybe one is the father of
the Spicer, here for example. Maybe one is the mother
of the Spicer, and so on. So we can explore that
a little bit further by looking at the trees and so on, of the individuals in
each of those clusters. We’re gonna get back
to the gray boxes here. But one interesting thing to note is, there a lot of people that won’t cluster closely with a particular group. So here, for example, we have a large group of gray boxes here. What you’ll note is though,
that they align here with the Spicer family this way, and with the Mullin family this way. And indeed, we can see, the Mullin and Spicers are married here. And in fact, these are
potentially descendants of that Spicer-Mullen marriage, which was far enough back in time where that’s certainly possible, there are many many
descendants of that marriage. These might just be other relatives of the Spicer and Mullin
families, particularly if they were married from the same place, maybe they had other
relationships and so on. Here for example, you can see these individuals
in this grey box here, they align with the Colwell family here and they align with the
Colwell family here. So they’re sort of in a way, they’re linking together
these two Colwell clusters. So, look at these gray boxes too, they can be really interesting in researching these family lines. So don’t rule those out. And in order to figure out
who those gray boxes are, often what you’ll do is,
you’ll just highlight them or hover over them, and
then when you come over you’ll see a box will be red here and you can hopefully figure out the name and find that person in
your batch tree and so on. Now, I can’t emphasize this enough, that clustering is not
proof, it is only a clue. You have to do all the
documentary research necessary to examine your findings. You’re not done when you
find a common ancestor, it is just the very beginning of the research that’s necessary to verify the the trees of the other
individuals, to ensure that the person you think took the
test took the test, and so on. A lot of additional work is needed, but these clusters are
incredibly powerful clues for our research. So please consider joining
the Genetic Genealogy Tips and Techniques group on Facebook, we’ve been discussing these
auto clusters quite a bit in the past couple of days. And if you’re interested
also in following along with additional and future
YouTube videos that I create, consider scrolling down and
clicking the subscribe button. Thank you very much.

16 Comments

Add a Comment

Your email address will not be published. Required fields are marked *