Easily Cluster Ancestry DNA Matches | Genetic Genealogy

[@FamilyHistoryFanatics]

🕵 Dos & Don'ts of Filtering Your Matches kzbin.info/www/bejne/nGW5p5RmjrSIg68

[@beowulfwheeler2580]

Even without endogamy, you can end up with confusing color combinations especially if several of your distant ancestors resided in the same area. For example your fathers grandfathers sister could have married your mothers grandfathers brother and thus the match can end up with both paternal and maternal clusters without endogamy.This happens frequently with me as both my parents ancestors lived in the same and bordering counties along the Hudson River for centuries without any common ancestors in recent history. if you then use one of these common matches to further sort without knowing a dual side match, it can be a disaster.

[@ValorieZimmerman]

Pedigree collapse and endogamy are always gonna futz things up, for sure

[@FamilyHistoryFanatics]

You make a great point. When you have a cousin who is related to you in multiple ways, you often have to ignore them initially until you've fully researched everyone else.

[@FamilyHistoryFanatics]

Valorie... that's correct. I made a video about the topic here kzbin.info/www/bejne/jZ3UkJRjnMmZb80

[@jessikamoore5033]

I have a bunch of this. My parents also seemed to be 4th cousins:(

[@mattpotter8725]

Read this and thought it was very much like my clustering. If you think this is bad you should try researching the DNA matches of a close relative (my 100+ year old grandfather) from a rural part of Ireland. Many of my matches are fine and just have one line (many who I think are from common ancestors who emigrated), but there are still quite a number who are in groups from both sides of his family tree (paternal and maternal), who I put down to those that remained in the region he was originally from longer and so closer back to the common ancestor. What is interesting now is Ancestry having recently added their paternal-material label to the matches. I'm not totally sure how accurate is or how Ancestry have calculated this new feature, but it is another thing to possibly use to verify a match that has say 10 matches in a paternal group but 1 in a material one. I think probability and caution are still needed to be used in assessing them though.

[@chrisferraiolo1935]

Here is my method for sorting Ancestry matches. I found it is best for me since it makes sense and is very easy. Check it out: 1. Separate the colors into warm and cool. Warm being colors like red, orange, yellow etc and blue being blue, purples etc 2. Create groups for each of your 2nd great-grandparents if you know them. If not, you can still make groups like "001" etc. 3. Assign the dots to each parent. There should be eight each. A first cousin would also have eight. 4. Build the tree down to matches or do a Leeds chart to make even more dot combinations. Look at shared matches, too. That way you can add more people to each group. Make sure the colors match on the leeds chart. 5. At the end of the sorting, everything should be trickling down on the match list. 4th cousins should have at least 1 dot. Hope this helps someone!

[@FamilyHistoryFanatics]

This is a variation on Devon's method. kzbin.info/www/bejne/pH2ul4WvoLZgfas

[@LadyM1966]

Chris, that is interesting ... I think I need to "draw" it first!

[@jeannettenemon3882]

Thanks for that, yes, hopefully I can add closer cousins, pending availablity! I also want to ask please if such a kit can be loaded for a DNA account with the two or so sites that allow free DNA transfers?

[@FamilyHistoryFanatics]

If you're asking whether a Superkit can be transferred off GEDMatch for use on other platforms, sadly, the answer is no.

[@jeannettenemon3882]

@@FamilyHistoryFanatics That's unfortunate, but thanks for letting me know. I hope some day the system becomes sophisticated enough to allow the creation of kits for those deceased family members where a kit is desired.

[@truthhurts5158]

you say that but what do you do with them after they are clustered

[@tammysue2334]

Thanks for video I never know what I am looking at lol 😂 I get overwhelmed ty

[@FamilyHistoryFanatics]

I'm glad I could offer some insights. Just remember, genealogy takes time. Pace yourself and it will be so much fun.

[@jeannettenemon3882]

For creating a Lazarus Kit, group 2, how distant a cousin can be added into the group? Can one place a 3rd or even 4th cousin in group 2?

[@FamilyHistoryFanatics]

You can place 3rd and 4th cousins in Group 2. To get the best results, try to add 2nd and 1st cousins as well.

[@staceycoates1418]

Well, my first comment is that I use the Mary Hill color scheme for my dots for the grandparents of the kit. Then I use the lighter hue (maternal) and darker hue (paternal) for the grandparents of the grandparents of the kit. That takes up 12 of the colors. I also do use a dot color for a group for people in my tree and people not in my tree. If Ancestry ever gets us a way to filter by that then I may get rid of those two groups.

[@FamilyHistoryFanatics]

Devon briefly mentioned Mary Hill's color coding in this video. It's not DNA but it's related. to organizing. kzbin.info/www/bejne/onmrnWOCqsqsetk

[@wendys1404]

Something else that is helping me more than the colors (until I learn that a little better).... With the notes for that match, I will look at the matches in common and the notes I put on them. If there is a common ancestor, I will put that in my notes for that match. That way it lets me know what family quicker than having to see the group color for them....

[@FamilyHistoryFanatics]

Great tip. I've been traveling and missed it. Thanks for sharing!

[@LadyM1966]

Still watching, but for me I can't do Paternal/Maternal clustering as shown in the video due to pedigree collapse and/or endogamy in my family. I have a paternal 1C1R. When looking at our shared matches, some are on my Father's side, some on my Mother's.

[@FamilyHistoryFanatics]

Endogamy and pedigree collapse do complicate things. kzbin.info/www/bejne/jZ3UkJRjnMmZb80

[@GreenBananaz]

Hi, I’m so glad you posted this. I wondering If anyone could please help me. I use a colour coding system to group certain lines together. My mum has taken the test so has her half brother. They share the same paternal mother so I’ve coded all those matches yellow because I know that’s the WEBBER line. I found a 2nd cousin who I connected to my mum through my grandfather mother’s line MASON & this was confirmed also through talking to her on Facebook so I used dark red. There were still matches with my mum that neither the Webber or Mason line were linked too so my presumption was they must be my mothers fathers descendants because I have no record for his father so I used green. I still quite have a lot of matches with my mother that aren’t linked to any of the other 3 colours. Could they possibly be my grandfathers grandparents on his fathers side? I used orange to group them. Also within those certain matches there are groups still coded orange but not connected to the other orange groups. I’m so confused 😂 thank you In advance. I hope what I’ve wrote makes sense

[@FamilyHistoryFanatics]

What you've written makes sense. I'd love to tackle this in a live stream Q&A. Would you share some screen shots and further information with me? You can start the process by using our contact form www.familyhistoryfanatics.com/contact

[@JustAGuitarPlayer]

I love organizing my dna matches into groups but the limit of 24 groups on ancestry isn’t enough. I have used up all the 24 groups (I have eight 2nd great grandparents couples and sixteen 3rd great grandparent couples but many of my dna matches and I actually share 4th, 5th or 6th great grandparents as our MRCA so to have to put them into one of my great grandparent lines that aren’t really their shared great grandparents is frustrating. More than any other feature, I wish Ancestry would increase the amount of groups we can create up to at least 124 groups which allow me to create groups for all great grandparents up to at least my 4th and 5th great grandparent couples groups. I don’t have any first cousins. Only 1st cousins once removed and 2nd cousins and beyond but for me to really figure out my tree, the ability to group dna matches further would help. Nice video on how to do this for those who haven’t discovered it yet but if you have any contact with ancestry, I hope you’ll pass on my more groups request. Thanks

[@FamilyHistoryFanatics]

While I understand the struggle, I know that Ancestry tried to use more colors in the past but the colors were too similar for people to differentiate between. So, perhaps they could use color and pattern combinations? We've offered that suggestion before, so send your feedback using their support link support.ancestry.com/s/article/Providing-Feedback-About-Ancestry?language=en_US They need to hear from customers and genealogy educators.

[@vancraven7688]

Why couldn’t ancestry set it up where an individual could import all kinds of colors and shapes,icons or multi striped colors?

[@turnerdan53]

I have a new match that is at the top of my cousin list. When I do the shared matches they show up as both branches. If both of their parents are 2nd cousins would that add DNA to make them look like they are 1st cousins?

[@FamilyHistoryFanatics]

That is correct.

[@mattpotter8725]

Great video, I wouldn't personally do the clustering for maternal and paternal (I think record based, or even just talking to living people should be able to get that), but start at matches of just below 200cM, add the first one (and so their shared matches into one group) then create a second cluster group for the next highest match below this value and add them and their shared matches into this group, and then rinse and repeat until you hit a certain lowest limit for cM that you have decided upon, or you run out of groups on the site (you've used all 24). For me having matches in multiple groups, even in both material and paternal ones, isn't a problem, it's just more of a challenge, although it all or a large percentage of your matches fall this way then it is, obviously. I would also say that although it's not directly related to actually doing the clustering, so maybe I'm ok with you not mentioning it, but at some point you will need to do the record based research and if you are interested in building a family tree then the first steps should be firstly talking to your living relatives would be the first step, and doing some record based research should be the first steps with this clustering only coming after you've done that (obviously if that isn't possible for any reason and you need a starting point it's different, but for the majority of people that isn't the case). I would love there to be a follow up video to this discussing possibilities of looking at any shared matches of shared matches (even shared matches of shared matches of shared matches) and how to use them, the use of adding them to these groups (going down to the lowest reaches of the match list) as that's what I'm doing right now. I've done extensive record based research and used the higher cM matches to link them into my tree (as much as I can at least, it's almost impossible if they don't have family trees and don't respond to messages), the low hanging fruit I've now picked, and I'm not looking to get further back, which means going more info the unknown where I think I have to work on probabilities of matches matching other matches, where sub groups have one or two members in known groups (grandparent clusters), and where everything becomes a bit blurry and unknown.

[@FamilyHistoryFanatics]

My wife constantly tells people to build their family tree using records so that I can focus on talking about the DNA side of things. So, I hope you don't mind me trying to focus on one thing per video. But I'll see what I can do about your video suggestion.

[@mickscrafts]

Thank you so much!

[@FamilyHistoryFanatics]

Any time. I hope you'll watch more videos that we have about AncestryDNA or genealogy research that my wife creates.

[@cruzinbosco]

I think it would be a bit wasteful for some to sacrifice any of the colored dots for paternal and maternal clusters. Maybe in the beginning it could work, but it's so much better, in my opinion, to just immediately mark them as one or the other using the "Add Relationship" button. That way you still have all your colors, and it's clear as day which side is which. I have to use dark mode which is not yet available with Ancestry (so I use a 3rd party) so unfortunately, I only get 7-8 colors as they all look like the same hues in dark mode. Therefore, those colors that I do have are very precious to me. 😂 But yes, this would be easier if you are just starting out and have colors to spare. It was a great tutorial. :)

[@FamilyHistoryFanatics]

Fair points. I did a video about labeling relationships which would explain how to do the Add Relationships you mentioned. kzbin.info/www/bejne/mnrSqHRoZ8t9Z7c

[@gregmiller238]

Just discovered this SO COOL. I adopted what do you do?

[@FamilyHistoryFanatics]

Start with this video and follow along the recommended videos at the end of each. There's a whole process. kzbin.info/www/bejne/l6KnY6Rvar5nnbs

[@aw4275]

I got lost after you did maternal and paternal side. Not sure how to get rest of both sides.

[@FamilyHistoryFanatics]

Do you know how at least a 1st cousin or 2nd cousin is related? Meaning, do you know if a 1st cousin is related through your mom or your dad?

[@aw4275]

@@FamilyHistoryFanatics yes for my dad side but not Mom side. There are only 2 people I know that is labeled mother side I know. Everyone else that has mother side I don't know.

[@FamilyHistoryFanatics]

Follow the steps to work through your father's side of the family until you can't. Those matches will begin your parental research questions. For your mother, make a LEEDS Chart to cluster your DNA matches. kzbin.info/www/bejne/Y2iXfX2vn9aib8U Using your father's known relatives, you can begin to group your DNA matches into clusters. With luck, you'll begin to see who is related to your father and who is likely related to your mother. You can then use the maternal matches to see who might have a family tree online. Look for overlapping ancestral couples. That's the general process. Here's another series of videos that might interest you. kzbin.info/aero/PLcVx-GSCjcdmsw25mbI-wJin_9_9QQUzI

[@pinwheelgrl9304]

If you have a true blue aunt/uncle or first cousin, you can use their match to "separate" your family into maternal/paternal dot-wise. Better if you have more than one for each side to use. But they have to be full-first-cousin or full-aunt/uncle in order to start everything out on the right track for dna purposes. First cousins will separate stuff into grandparent groups (2 piles). 2nd cousins will separate stuff further into great-grandparent groups (4 piles). Only go down to 90 cM. The smaller down you go - the more people will be related to both sides of the tree due to relatives in the far past. This is a great channel. I also recommend Your DNA Guide channel for a different way of using the dots to separate out branches. Her vids are easy to figure out too.

[@mattpotter8725]

​@@pinwheelgrl9304 Whilst I agree on the whole with your comment I don't agree with your comment about only going down to x cM, unless by this you are meaning to save time. I think it depends on the situation you are in (by which I mean how much research you've already done in the traditional records researching), but finding an ancestor that falls into a paternal and maternal group based on shared matches is fine, it just makes it more difficult to narrow down how they fit into your tree, so the other low hanging fruit I would address more. The annoying thing about Ancestry is that they don't tell you how much DNA the matches share with your shared matches, which might help prioritise groups which shared matches of unknown, ungrouped matches two over others, so say you knew this and your match has 10 shared matches, 8 which match a paternal grandparent, and 2 that match a maternal one. If you knew that those 2 had 150+ cM matches whereas the 8 were all in the 20-30cM range then you'd prefer them in the material. Without this knowledge you can't really know or might have to make assumptions that might be wrong (but kinda have to say some points, genetics in the end is a matter of probability, nothing is set in stone when you start to go back more and more generations, which is what most of us doing this are really after). Having said all of this sometimes you do get some small cM matches that only fall into one group, that have extensive family trees and that you recognise surnames that match the ones you have in your tree and with a bit of paper based research you can link them up, and by doing this, and messaging them, and hopefully getting a response, can find out the cM some of your shared matches with them share with your shared matches in common (I've been trying to help a match locate a common ancestor by doing this, though they have to be very interested in going further back on that branch of their tree, which they were very interested to do so in this case). Personally I think doing some record based research first is needed before starting on any clustering as above. By doing this I would say that creating groups for each parent shouldn't be necessary, and you start at the grandparent level. I think clustering is really only a tool needed when you hit brick walls and can't get any further (if you are adopted then maybe it's a different matter, you probably need this at the start, but that isn't the vast majority of people). In my experience I have 4 grandparent clusters, a handful of great grandparent clusters, and quite a lot of unknown clusters (some of which overlap with others, sometime both paternal and maternal ones). In the end there is no right or wrong ways of doing things (ok there are probably many wrong ways) but some approaches where you do need to be cautious as if you ancestors and their descendants didn't migrate that much then I think it's almost certain that in the majority of cases you'll have some matches that fall into multiple categories, but some that don't, and those that don't can be real gems imo.

[@alanheadrick7997]

I am using 23andme for this idea. I see a 3rd cousin that I think maybe on the father's side. She has only three shared matches. So if I understand this correctly those three matches would also be on the father's side, but what about their shared matches? The reason I ask is one of them has 92 shared matches.

[@FamilyHistoryFanatics]

Often after you consult your matches and their matches, you MAY find clues to help you figure out how the extended cluster is related. It's possible your matches' matches are related to you but you didn't inherit the strand of DNA. What you'll want to do is build their family trees to see if they intersect with yours.

[@alanheadrick7997]

@@FamilyHistoryFanatics Most of them are likely adopted and Chinese. I might be lucky to get info from maybe 5%.

[@alicemusiclive3406]

How can i create groups with one 859cm (12%) my biggest then i only got 3rd and 4th cousins just after ?! :/ I don’t have matchs with 200/300cM It’s 859cM then 67cM…

[@FamilyHistoryFanatics]

Without several documented DNA and paper trails, there is no evidence that you inherited any DNA from any specific ancestor on that line. You inherited on average 12.5% from your great grandmother, but by the time you get back to the 1100s, you would have thousands of ancestors for your great grandmother and you (and she) probably only have DNA from a handful of them.

[@glenjones6980]

I'm new to DNA results and finding at best just 2nd cousin matches. Based on the paper trial my ancestry is 100% English events for six generations on both maternal and paternal sides, but I'm not getting matches to either grandfather. It seems that at least one set of grandparents on paper is proved wrong with DNA. I have a huge number of unlinked/no tree matches and therefore very little to work with. With the help of a DNA savvy friend, we have managed to sort some basic maternal/paternal grouping but the lack of trees is causing huge problems as I can't say for certain who my grandfather is. My closest matches are mostly based in the USA and the few who have replied to messages just state we don't have a link. Clearly we do but the lack of available tree data is a problem, if they would just reveal the surnames at great grandparent level it might help but there is a wall of denial and silence, I feel as if they all know the secret and are refusing to let me in on it/

[@FamilyHistoryFanatics]

Yep. You've encountered a large problem in family tree building. I hope you have clustered your DNA using the Leeds Method kzbin.info/aero/PLcVx-GSCjcdmJ2ZbhDCBPugPMYveHkwxs Then, you can try to figure out who people are using what available information there is. Another thing you can try is descendancy research. This is where you will use your paper trail to find descendants of of each ancestral couple. (This will eventually involve researching living persons). Then, you would reach out to the living persons to see if they have taken a DNA test or would be willing to for you to establish the identify of your shared ancestors. It's not easy and many people who take DNA tests are only interested in ethnicity results. I don't think that's going to improve. But, there are ways to attack the problem. It just requires a little more learning. I hope I have given you some key topics to investigate.

[@pinwheelgrl9304]

People love to hang onto the family skeleton for some reason. If you are young, the people in question may still be alive; but if you are older, I see no reason why some information can't be given. Like Andy says, Leeds is the best, use the spreadsheet version. Dana Leeds just did a free vid on it over at Legacy Webinars, free until around Oct 20. She shows what to do with overlap groups and explains why they happen, which I found most helpful; then goes on to solve her problem. Also, don't forget (on no-tree matches) that you can look at the shared matches for trees for clues on surnames. Also, if you know an ancestor further back or married to the person in question, newspaper and funeral home obits are great for clues on who is who. FindAGrave is also nice, but not totally dependable if there is no actual grave (vs someone just uploaded a memorial because they couldn't find one anywhere). Use that to work your way back down, and don't forget using FamilySearch (the records part). Also, Diahan Southard has more ways you can use the dots to separate out branches from the others without any contact or trees. Her YT channel is Your DNA Guide.

[@DeLaTr0ll]

Pedigree collapse was discovered on my partners side. It’s disappointing and very hard to figure out who is who.

[@FamilyHistoryFanatics]

Pedigree collapse does make the genetic and genealogical research more challenging.

[@ValorieZimmerman]

With pedigree collapse I don't worry about sorting & dots so much -- I just put them in my tree if possible. Ancestry is probably never going to be able to tell us all the various ways we're connected to such people, but we can do it "by hand." It does take a lot of research tho! On the other hand, there are fewer ancestors to research.

[@roberthopgood1894]

i am still confused

[@PaulJHawkinsJr]

Lol I have a Jennifer Kevern in my match list.

[@FamilyHistoryFanatics]

Awesome.

[@scentials]

I think 1st half cousin would work the same as 2nd cousin.

[@FamilyHistoryFanatics]

It really depends on what you want them to do. They can help you segment from one line from another. But, they can also create other problems. So.. try it out and see what happens.

[@pinwheelgrl9304]

Half-relations mean the MRCA for them and you descended through a different spouse. So that match can be used to single out that one line, vs "the couple". In theory, yes, that person actually gets 'bumped up' a line, so they are very useful if you can't find a 2nd cousin for that spot. That's what I use mine for. But Andy is right if there are special situations.