Sub-Clustering Your Shared Matches at AncestryDNA

[@christinaedwards989]

I manage over 20 kits between 4 generations for 6 people on 5 sites. This is going to be 100% helpful. I have color-coded me, my sister, and my son, during this process; however, I did not look at the matches. *I will go to each of our accounts and do that one by one.* I already know that some of us bring people in that others do not. I will be able to create a list and take that list to each of the other sites to locate more matches that will go into these subgroups. Super excited. Thank you!

[@tracygregory5293]

Thank you for all you "how to" videos. They are so well organized and presented. It is so hard to learn from a talking head without screen shots. You include every step and before and after screen shots! Thank you so much!

[@steveschulte8696]

This process works. It helped me find additional cousins in the 3rd to 4th cousin range, and to research the common ancestors of all of us. It has lead to "brick walls" that will fall if I devote some time to deep research. It also leads to clusters of people for whom I have few close relations among close cousins.

[@lmaniac58]

Thank you for sharing all your DNA knowledge with us. I always find it helpful and love your webinars. Thanks again!

[@BlaineBettinger]

Thank you for the kind words!

[@MrBookcollector]

First filter your matches by common ancestor hints. Add the common ancestors to your notes field for that match. The CA will now appear when you bring up shared matches. Create groups for the CA and add the shared match to the group if two or more CAs appear. Remember you are working with hints. They are very good hints when multiple shared matches agree on a CA. Then enter the CA(s) to the match you are working with.

[@elainewilliams8666]

Thank you. I love how your mind works! If only I had the numbers you are working with. I am in Australia with mainly British forebears, so i am slowly plotting the cluster groups together. My good DNA matches are the ones that emigrated to USA and Canada and especially those who have LDS as their religion, they seem to be the group I am matching with the most.

[@abudebbie]

Hi--thanks for this! My issue is that I have assigned colors to lines in my tree already and ancestry only gives you 24 colors to use. I don't have enough left over to do this unless I reassign people already color tagged, or throw out the color tags I've used. Is there a work around?

[@ruthwilson9449]

I have this too. I made lots of little Ancestry groupings trying to make some sense of it. I tried using the Leeds method in a spreadsheet, and it gave really interesting results.

[@julia2jules]

Thanks for this Started my ancestry dna two weeks ago I had started sub grouping just like this. I wondered if it was going to be helpful and now I know it will be

[@BlaineBettinger]

I hope it works very well for you! Good luck!

[@junepeasgood1907]

This information is brilliant, I watch Christa Cowen and I don't think she has one that covers this - because I have been doing matches that hard way. I'm now going to check for videos on where you put your "built trees from sub lists" - as I have several trees relevant to individuals who have very small trees (3 or 4 people) or who have quite a lot but they are all deemed private so it's hard to work with, especially when they are not in the UK. Found this video very useful - thank you :-)

[@karendasero6100]

Thanks for this tutorial. Your sharing of knowledge and processes are so helpful.

[@gavanwhatever8196]

So you've used 12 of the possible 24 custom groups doing initial organisation on one branch of your family tree. What do you do when you want to investigate another branch? Delete all groups and start again?

[@barbarabird3827]

When do we get the video for endogamy & multiple 1st cousin marriages?

[@marbah1889]

Wish I had that many in common match’s to work with! Great tips! You come as a match for my match, we could be related.

[@BlaineBettinger]

Just a year or two ago, this wouldn't have been possible. So hopefully your matches continue to increase quickly and you'll be able to do things like this soon. (Hi possible cousin!)

[@steveschulte8696]

I wish I had the problem of few matches. I have 938 4th and closer and 41,000 total. I have found family matches below the 4th cousin threshhold.

[@mikeposnick2376]

It seems to me that the methodology proposed by Diahan Southard is easier and makes more sense than the method you advocate for clustering.

[@pennyanderson2153]

I also have an Unknown #3 just as in your tree. Did you identify your #3? If so, how did you do it?

[@annisaacs1132]

Question: How close is too close for a match to be useful? I have two H1C matches that are from a known line. The first (closest match) has 497 shared matches. I started to mark them, but before I spend the next several hours going through them, I wanted to see if they are mostly shared with the second H1C. It's about 1/2 and 1/2 shared matches. Should I be marking these matches as a subgroup even though I know these two cousins are descended from the same grandparents? Or should I just skip them altogether and go to a more distant match?

[@jayfletcher5096]

I have the same question

[@lauragray19]

Same question. I did this with an uncle and his first cousin, and the initial IVW group was over a thousand people!

[@jenniffermichael]

The + Add/edit groups is not immediately available. click EXTRAS at the top and then choose Ancestry lab. On this page select enhanced matching on? otherwise, you will be pulling your hair out like I did

[@BlaineBettinger]

Great point! You have to be in the matching beta, which is available to everyone as Michael describes.

[@andreashiock301]

@@BlaineBettinger Adding "new groups" is not available for me. I have already grouped most of my matches and I now want to sub-group them the way you described but I no longer have the option to "create or add new groups". I checked Michael's suggestion but it tells me there are no Beta's available. How can I add new sub-groups now? Thanks so much. This feature is no longer available to me: In the top-left corner of your list of DNA matches, click Groups and select Create custom group.

[@lindavies9948]

Thank-you, this was very clear and helpful. I have already clustered my groups slightly differently but your approach is interesting.

[@jlowery000]

I have over 130K of matches. Does this method work well with that many matches. I have several branches that stop at the 3rd great-grandparents and one at 2nd great-grandmother. I have watched the vid 2 twice and think I understand. I went in right off the bat and started putting Ancestors in family groups and marked them with the colored groups. Now I have only one spot open. Blaine any help to this DNA dummy would be appreciated. All the vids I watch never have the amount of matches as I seem to have.

[@jwebbtx1961]

Am using 1st cousin for first cluster of Shared Matches....down to 25cMs and already have almost 850 matches. Looking for GGGrandmotner. Thinking I'll find connection in lower cMs. Love this video and have now ordered B.Bettinger's new book. Hope my wrist holds out!!

[@cecelliarogers7346]

If you have 2 shared matches, and you know one to be the child of the other, do you treat the child as you would a close relative and just leave him or her out of the subgroup?

[@BlaineBettinger]

GREAT question!! I may include them (to avoid re-asking myself that question every time I "re-discover" the child belongs in the group), but I would likely put a note in the note field indicating something like ("John Doe - child of shared match Jane Smith (ignore) ").

[@mnrd34]

I'm having fun doing this, even tho I have 748 vs your 150. My problem is in wanting to go lower than 20 cM for matches. I have some saved from before Ancestor took away everything under 8 cMs. Would love to incorporate these saves, but can't find a way - and don't want to put them in the wrong group (not being 100% certain that I have them grouped correctly, to begin with). BTW - thanks for ALL you do and share.

[@barbarabird3827]

I get it: I need 4th & 5th cousins to help me figure out who a 2nd gtgrf really was - my closest cousins (3rd) likewise have found 0 links for him, using the name he went by. However I do have what seems to be an extended ghost family -seemingly at appropriate cM distances. 4 names keep cropping up, 3 of whom regularly intermarried- never find one without at least 1 other, - & all tended to marry 1st cousins. (I need a video on this!) When you get to this situation, it's worth doing a search for each of those names. Below the 20, I call it digging in the cellar- genetic archeology. Aside from the potential tree info, it CAN be useful to click on their shared matches - there may be some! It does work in that direction , if /when matches exist. (I wouldn't worry about filing anyone in wrong group - click, they're out of it - click, into a new.) Good luck!

[@IftheShewfits]

I just subscribed to your KZbin channel! You offer so much helpful information! Thank you! I'm wondering if this method is essentially the same thing as the clustering report one can generate through Genetic Affairs?

[@BlaineBettinger]

It's very similar, yes.

[@christinaedwards989]

This is great. However, I don't have the all matches option while in groups. ??

[@jayfletcher5096]

I'm trying to find common ancestors of a great-grandfather. I've started this process by finding matches between myself and a second cousin once removed. However, I have several second cousins related through the same great-grandfather, and they show up at the top of the list of matches with the my first matched second cousin once removed. Should I exclude these other second cousins once removed when I start making subgroups? It seems like they will muddy the water.

[@darylcoleman5342]

Blaine (or anyone), I'm stumped. When it comes to the 1st Sub-group (and any subsequent sub-groups), the option 1 and option 2, you choose to go with option 2. I'm not clear regarding the difference in the results between the two options. I can see from already trying that the lists are different, but I'm not clear which option is truly more helpful. If this could be fleshed out a bit more, I would be appreciative.

[@steveschulte8696]

Clustering with option 1 leads to smaller sub-groups of closely related persons. All matches share DNA. Clustering with option 2 will lead to clusters that share DNA with another sub-group, that is from your grandparent, but may split the groups at the great-great grandparent level.

[@8lindamae]

Thank you so much! I was stuck at the initial cluster grouping stage and didn't know where to go from there.

[@tarasmonax2595]

Has Ancestry changed their format? I don't see the All Matches tab shown by Blaine Bettinger at 11:33 in this video. What am I missing?

[@ceecee55417]

I have the same question. I also do not see all matches.

[@kparhow]

Blain, here is a question about the effects of a pairing between first cousins. My spouse has a marriage between first cousins a couple of generations back. After several miscarriages, we had our DNA tested and my spouses came back with a balanced translocation between chromosome 6 and 16 (non-standard karyotype: 46,XX,t(6;16) (q21;q23)). The interchanged pieces involve portions of the long arm of both chromosomes. The same number of SNPs are present, just some are moved to another chromosome. A sibling also has his translocation. We were fortunate to have a child who was tested and has a standard karyotype. The questions are: How do the different Autosomal DNA testing companies (Ancestry, 23andMe, etc.) handle the translocation when they process a DNA sample? How does this affect the DNA Matching results? Thank you.

[@staceycoates1418]

My issue with this is that I already have many of my colors taken. For my specific matches I have tested my all four of my grand parents so I am almost immediately able to tell what quarter branch, and with my grandmothers, what eighth branch. Because of that I don't have many colors left and I am not certain how to best utilize them. Do you have suggestions on that?

[@BlaineBettinger]

Unfortunately, I do not. Perhaps provide constructive feedback to Ancestry asking for more colors.

[@darylcoleman5342]

About a week ago, when I was still actively assigning colors to surnames, I requested to Ancestry that then up the number of colors to at least 32. This morning I started working on Blaine's method, and because I needed colors, I started with the lowest used colors and re-designated one by one for use with this 'new' method. So far, so good. Besides, this new method Blaine is showing in this video may in fact be more productive than what I was doing previously.

[@lindavies9948]

Stacey Coates If you are managing all 4 tests separately I would think you could do a separate tree starting from each grandparent and going backwards. Then you can use the coloured dots for the subclusters for each tree independently.

[@abirdchirps6890]

Thanks for sharing and teaching tips! This is awesome.

[@cassv350]

Funny I figured this out before viewing this already, but good to know I am on the right track

[@barbaramythen5774]

Working on a paternal line (maternal line is known) but will this process work for a half cousin who is top match on paternal line?

[@MusicInMotion_67]

How can you do this with just your account? If you don't have access to other people's accounts?

[@CarolAnn61]

I’d love to understand this but I’m completely lost.

[@raftisa]

Unfortunately, I can't hear your video...is there somewhere else I can obtain this information. Thank you.

[@MusicInMotion_67]

I am trying to find my paternal grandparents. I know my paternal GG possibly 2GG (haven't figured that out yet) a) does this help with narrowing down grandparents and if not is there a method that would? b) how far down should I start? So far my closest match is 567 cM and I know who our GG possibly 2GG are.

[@lauriepringle9303]

I am trying to get my head around this. I haven't a cousin, but I have a niece. Will this skew results, being off by a generation?

[@CoNaana]

Any tips if you suspect multiple common ancestors?

[@christinemulder8661]

Hello, I am trying to follow along using the Ancestry matches that I have already started sorting into color coded groups. I may have started off wrong here though - When you start with two cousins, I assumed that one of the cousins was going to be you, ie, the DNA tester. And I realize after trying to do this, I may have been mistaken in this assumption. Who are the cousins that you start with?

[@BlaineBettinger]

You can use any two cousins, as long as you have access to at least one of the cousins' DNA. I was using my father and his 1C, via my father's DNA.

[@RoyPounsford]

Thank you a clean understanding how to approach shared matches

[@BlaineBettinger]

I hope it is helpful!

[@joebissett1]

Have you refined your process for FTDNA and MH matches? If so, could you share?

[@BlaineBettinger]

I don't have a similar process, yet, for those shared matches. I manually review shared matches there.

[@jacquelinehunter488]

Perhaps someone could clarify please - does this method only work if you manage more than 1 kit?

[@BlaineBettinger]

Although I had access to both accounts in this example, it wasn't at all necessary. This works just fine if you only manage one kit.

[@joebissett1]

I'm cleaning up previous attempts at clustering by now using your method. I now have an old empty group name, and I can't find any way to delete it. Any suggestions?

[@BlaineBettinger]

Go to add/edit groups next to a match and click to show the drop-down menu, then click the pencil (edit) icon next to the group you want to edit.

[@joebissett1]

@@BlaineBettinger I'm embarrassed that I did not find it. :-( Thank you.

[@BlaineBettinger]

@@joebissett1 no worries at all!

[@rondascroggins2477]

I have too many groups. Is there any way to download this information .csv, .txt, whatever so I can put it into an Excel document? Then I can cluster better.

[@2011jdp]

dnagedcom.com can download your Ancestry matches and you can export it to excel I believe. Requires a minimal fee.

[@SouKube]

Great videos! I really enjoy it. Do you have one that shows raw data from the dna co.s . I was trying to understand how are the sequence matches done from the data to parents, siblings, cousins, etc. I mean do they take the field that shows the AT GC etc. And make a long strand pair - one of the 1st letter and another of the 2nd letter in each chromosome and then look for parts of it that matches the same chromosome in another person's data. Does it have to be the same chromosome. It wd be good to see how mom/dad matches share 50%, or the percentage match in sibling/cousin/etc. Do you know how this works. Love to hear back or in a video. Cheers and thanks for all the stuffs u r doing.

[@luannpatterson6258]

I've been practicing with my own family. Is this process in print somewhere? I tried making notes but I still get lost. Hoping to figure it all out & finally get a mystery person solved.

[@totopcat]

Wish I had known to omit close family members at beginning. I was following your actions along with the video

[@BlaineBettinger]

Thankfully it only takes a few clicks to remove them!

[@2011jdp]

Would it be a similar effect to just star all the matches for the highest ICW match and run an Autocluster on geneticaffairs using only the starred matches?

[@mattpotter8725]

I know you posted this message a while ago, but I've only just found this video and your message and wanted to respond (you've probably tried this and other methods). I thought this might be useful for others (and possibly you) to know. The problem with this approach is that the highest ICW match is the highest match to you, not necessarily to them (which you can only know on Ancestry by messaging them and asking, and it assumes they respond and are willing to help) and so by doing this you might be missassigning them to a certain group. They certainly belong in this group/cluster, but your might be excluding them from another group that they might also need to be in (and adding them to both shows the two groups are from the same area of your family tree, which might help identify where each of them branch off.

[@loiscampbell3510]

LOVE this

[@BlaineBettinger]

Thank you!

[@cynthiahess7477]

This information was so helpful. My question is what ancestor should you choose to start with, ie should it be a first cousin, second cousin, third cousin? I was fortunate to find about 20 cousins of various degrees through Ancestry DNA, we are all descended from six brothers that are mostly our great grandfathers. We are trying to break through that wall to confirm our great great grandfather and find our third great grandfather. Thank you so much for this information, I can’t wait to pass it along to my Armstrong family to start this process...

[@BlaineBettinger]

It depends on your brick wall. Since you are working with a great-great-grandfather, you would ideally work with one or more third cousins descended from that same great-great-grandfather.

[@cynthiahess7477]

Thank you so much for your response. After making all of my sub groups I was hoping there would be a way to print them to kind of compare side by side with all of the color coding to really dig in to follow the lines...but Ancestry does not include the coding established other than that specific subgroup. Can you recommend a charting program to plug these subgroups into to be able to print out and share with family members. I found Progeny Genealogy charts online, but not sure that is the direction I should take.

[@mandlerparr1]

Right now, a good place to look is wills, if you haven't already. Also. location research. If you all suspect that your great grandfathers are siblings, look for the will of that suspected 2x. use the card catalogue, because many wills aren't indexed. also search on familysearch. Is it too far back for census records to help? look for probate and guardianship records also. Also, if you all have a suspected 2x, put him in each of your trees and see what comes up on your Thrulines. Pardon if you have already done all this. Do the same for his suspected spouse, just make sure you are all putting the exact same information.

[@cynthiahess7477]

@@mandlerparr1 Thank you for the great suggestions. I also thought about searching all the cemateries of the town we believe they passed away in, and searching the surname. The problem is that Ancestry is giving us a suspected famous general with the same last surname, but all research shows that his children were definately not our relatives, so we have to find that link, as maybe they were brothers or cousins, or something to that effect.

[@Rileygenerations]

This was an AWESOME tutorial.

[@tonycureington4440]

Thanks! Very helpful!

[@jeffjenson111]

Keep crushing it. Nice work.

[@BlaineBettinger]

Thank you very much for the3 kind words!

[@alanheadrick7997]

Strange I checked for shared matches and there are none. I know one pair is a mother daughter, so they should be shared. 23andme works great for this, Ancestry not so much.

[@BlaineBettinger]

There's a threshold for shared matches at AncestryDNA, where the shared match must be an estimated 4C or closer (i.e., greater than 20 cM) to BOTH matches. A threshold is essential, since shared matches with less than 20 cM have an increasing chance of being shared via means other than the shared match cluster (that is, shared on multiple different lines).

[@kylecook2437]

@@BlaineBettinger So, are you saying do not group matches less that 21cm?

[@kayldee1215]

I must be the dumb kid in this class....... if I have 136 with my first my first match and go all the matches would i not need 139 colors?? I know that is not the right answer, but I remain completely and utterly confused. H E L P :-)

[@lindavies9948]

They won't each be in separate clusters, the clusters will have several people in. As you go down the individuals you will be adding them to clusters you have already created.

[@jaclyndouglas8471]

Thanks for this. Although because both my parrents are only children this didn't work as well as I was expecting.

[@paulaalexander7521]

What clustering sharing?

[@kathrynludrick4821]

Awesome

[@DS-zj2on]

Thanks!

[@Lindys]

Very interesting. Thank you.

[@MsToboo]

Blaine, I love your work! I think this will be useful to find my unknown maternal grandfather. I have my paternal grandmother tested, and my paternal grandfather's half brother, so I can exclude all of those matches. On my mother's side I have her half sister tested, so can exclude any of these shared matches. As I'm in Australia, this honestly leaves me with 12 matches that don't match any of these known family members' kits, and I suspect some of these 12 also match one of my 3 known grandparents, they just haven't replied to my message, nor have a tree with any information. I know it's a waiting game until one of these people reply, but with such a low number of possible matches to my mystery grandad, could this imply endogamy? I do know he had Huntington's Disease, so if he didn't have any siblings without Huntington's Disease, I'm potentially looking for a low cM match aren't I? PS - I'm looking forward to seeing you in Brisbane at the DNA Down Under roadshow! #DNADownunder

[@RoyPounsford]

Blaine, You need to keep it as simple as possible.

[@joeflood8232]

Sorry Blaine but you could have described that in 60 seconds. Too tedious for me to get more than half way.

[@kathleenthompson6007]

Can you use this to find unknown father?

[@BlaineBettinger]

Quite possibly! If you can cluster your paternal matches into one group (such as by testing your mother or close maternal matches and taking all the close matches that do NOT match your mother), and then sub-group them, you may discover discrete sub-groups that point to different paternal relatives. Then you can find how those different paternal relatives intersect to point to your father.

[@kathleenthompson6007]

I found a half sister but I need to prove it to her. The thing is I connect to her surname .

[@mandlerparr1]

I did so without this, I had to build trees of matches in common (so, if they were in common, I put them in a tree and researched until I found the connection). Some people use excel to group common matches. This makes it much easier to group them. But, after you get them grouped, make a tree for each group starting with the strongest match with the best tree. Or the one that looks like you could build a good tree out of it. Start adding matches to it from the subgroup, building the tree out for each match as you go, until you find the connection. For this type of tree, you will want to include siblings and children, don't go straight back only. Go straight back at first a few generations, but then start doing descendant research if going straight back doesn't lead to a common ancestor right away.

[@kathleenthompson6007]

mandlerparr1 I found a half sister at 1948cM but she is in denial. I connect to .her great grand parents on a different line. So if I put then in a tree would I be out of line.

[@mandlerparr1]

@@kathleenthompson6007 It would not be out of line because they are your ancestors. You cannot help how the actions of people in the past are affecting those in the present. You did not create the situation the two of you are now in. I find that in these situations it is best to just leave them alone for a while so they can get used to the situation. If they have asked you to not message them, don't. I have matched closely with people who have asked me not to message them and I don't. They are, however, in my tree. You can even put her in your tree, she will be private, so no one but you will be able to see it anyways. Your tree is yours and you can put anyone you want in it. Sometimes, out of respect, I will make a person private, but this is usually only in the case of children who have recently passed on. I don't want grieving parents messging me.

[@katemooers4147]

Thank you Thank you i was about to make excell sheet

[@BlaineBettinger]

I'm all for avoiding Excel whenever I can!

[@kjw79]

Hee hee sub-gloops :)

[@dorasmith7875]

Blaine Bettinger needs to quit muddling his hapless brain and trace trees.