Pages may contain affiliate links.
See my disclosure page for more details.

The New GMP Feature That's a Game-changer

Yesterday I had the time to get the most recent update to Genome Mate Pro (GMP). It has a new feature for AncestryDNA analysis that took it from useful-for-an-advanced-user to a tool I'd recommend as soon as you're ready for it.

This new tool allows you to see the "shared matches" (also known as ICW meaning "in common with") between more than two people. Let me explain that another way. On the AncestryDNA site, the list of shared matches is a list between two people, the test taker (the test you're looking at) and a selected match.

This is the most useful tool at AncestryDNA. About on par with it are family trees. (I give shared matches one-up because trees are a traditional tool that you can use without DNA. If you're interested in DNA, the shared matches are necessary, although you need to get trees in some way, too, to make it useful for finding new ancestors).

That means this new feature in GMP takes shared matches farther than is possible on the AncestryDNA site.

In this post I'm going to explain why I find this such a big help with using GMP versus the features it already had.

This is important because not everyone uses DNA in the same way---and we shouldn't use it in the same way. Our different interests as well as the differences in the ancestry of the test-taker mean different processes will work better for different people (and even for different problems). As with all genealogy, differences in how our brains work also make different methods or processes work better or worse for different people.

That means it's best to learn (briefly) about a few different ways to work with DNA so you can find what is best for your situation.

So read this post, see what I do. Then decide if you like it and want to try it, maybe you want to just take parts of what I've described, great! If you decide this is a total mismatch for your interests and needs, I highly recommend making some notes somewhere about WHY you think this. Then you can save time later by finding what addresses the issues you identified (making notes whether you love my ideas or not is a good idea to help you start to define what you need and want when you work with DNA).

About Genome Mate Pro

This post is most likely helpful for people not using GMP or who aren't actively using it. I want to start with what GMP is and my experience with using and not using it over the last few years. If you're an active user you can probably skip to the next section.

Genome Mate Pro is a free software program. Like many apps and programs it was created by a hobbyist genetic genealogist who needed a tool for herself (oh how I wish I had learned to program in college so I could create my own custom solution!).

That's important to know as far as getting support for the tool. This is not put out by a mega-corporation who has minions available to help you with everything.

GMP has a notoriously steep learning curve. I tried using it twice in the past (I believe in 2016 and prior to that, maybe 2014 or 2015 but possibly earlier. I had infants during that time so the whole first half of the 2010s runs together!).

I found the learning curve extremely steep back then. I don't find it that bad, now. I don't know if it's because of my expanded knowledge about genetic genealogy (I wasn't a complete beginner even in 2014). It might be because of another change.

GMP is now much FASTER to setup. It was more the set-up time than the learning curve that stopped me before. I want to mention this because you may have tried it and had the same experience.

It's possible the learning curve doesn't seem so steep because I can use the program more now because it's so much faster to get data into it (FYI, this post is not a "how-to" on setting-up GMP so I'm not going in-depth on how set-up works).

So that's why I am really "using" it this time around.

So what's the point of GMP?

It's about managing your DNA data and keeping track of what you're doing. In part it does this by providing analysis tools. In the past I really wanted to use GMP for chromosome mapping but now DNAPainter is available.

For me, the analysis tools are for "light" analysis. This doesn't mean they aren't powerful, just that it isn't the only analysis tool I'd use. I prefer DNAPainter for chromosome mapping and I use my 4 Buckets Technique for a lot of my analysis.

Others use GMP for much "heavier" analysis (light and heavy are fairly artificial terms in this case but I don't want to write a paragraph to badly explain what I mean, you get the idea that you have options).

GMP is a database so it also provides some organization. I would not call it an organizing software because if I designed an organizing software, it would emphasize some different things. But, that's part of why I think a post like this is helpful.

The difference in what GMP puts emphasis on, versus what I'd put emphasis on, is important to understand once you are doing this level of "DNA work."

GMP is chromosome focused. This is what I've always seen emphasized in advanced genetic genealogy work.

This is what I eventually discovered was the problem for me and the tests I worked with.

This is important, this difference is about the ancestry of the test takers. There is nothing wrong with the emphasis on segment data. This is at the heart of how DNA helps us with genealogy.

This is a difference you need to recognize, though. If a chromosome-focused method isn't working for you, it may not be due to your personal needs or interests. It might be due to the actual DNA. I'll explain this more in the next section but I wanted it to be clear that this is not a judgment of good or bad. It is a valid difference.

(There's an easy analogy I can make that women will understand. This is like picking makeup. With different skin types and skin tones, most cosmetic products won't work for any one-woman. They can actually do the opposite of what they are meant to do, even. A woman will have preferences but she also has to deal with the skin she has, regardless of preferences. That doesn't make the "wrong" product bad, it is just the wrong color or wrong formula for her specific skin. It's why the cosmetics industry is so huge. Thankfully I don't think DNA tools need to come in as many varieties so you won't be faced with the equivalent of aisles of moisturizer or foundation to choose from when picking DNA tools!)

So GMP is a tool that helps you manage and work with your DNA results by uploading the data from either the testing company or via a third-party tool. I love it because you can connect tests for the same person from different sites. This allows you to bring in what you've learned working with results at another site and apply it to the new site and all those additional matches.

GMP offers some organization of notes and emails (as well as a standard email template) in addition to the data tied to the matches you've loaded. That means it pulls all those things into one place.

My (personal) biggest issue is I do a lot of my own research on a Chromebook in the evenings so I can't run GMP (natively) on it. I know there are workarounds but they are too cumbersome for the brief amount of time I usually have for my own research. If I did all my work on my desktop, I would rely on GMP more. I use it as one tool for DNA work instead of a hub, which it could be.

There's also that chromosome-focus that doesn't work for me.

Not all DNA is the same

We all know everyone's DNA is unique but turns out working with it can be "unique," too.

The way I visualize the difference in what you need to focus on, based on the DNA, not your preferences, is as if there are layers to DNA work.

Everyone needs to do the top layers. That's a general review of matches and building trees. This is what I've put into a process in my free course, The Road to DNA Success. These "required" top layers should be the bulk of your DNA work. The general review isn't so time-consuming but building trees, is. Without trees, there's no point doing anything else because you have data, not family connections.

Next are a lot of optional layers depending on where you've tested and the matches you get. A lot of these layers you can pick and choose from based on personal preferences, not just because you "have" to get into them based on the test you're working with.

At the bottom is segment data which can confirm the DNA relationship. Here's the thing, although this confirms the relationship, you can still confirm a relationship the traditional way, with paper records or oral history. That means this layer is technically optional.

It is very hard to be sure a segment comes from a particular ancestor. Remember, part of the issue is you may have gotten a segment from grandma but she got it from someone in her ancestry. That means matches on that segment (once you even confirm if it's the maternal or paternal chromosome) could be a close relative sharing grandma, a less close relative sharing her grandparent, or a very distant relative sharing a more distant ancestor.

Knowing a match belongs to grandma's branch of the family tree is not the same as confirming a relationship based on the shared segment.

People often act like it is and this is where I've found the big difference in using a chromosome-focused approach, or not.

Note: Does it really matter how you find a new ancestor or prove a new ancestor? I don't think so. You may want to confirm everything with DNA. That's your choice. I'm happy to grow my family tree using any valid method.

If you've read other posts on this blog about DNA, you likely know my family is all from Georgia (the U.S. state). I mainly focus on using my great-aunt's results because that is the part of my family I'm most interested in researching. Her grandparents were 2nd cousins. There was a lot of intermarriage between the same families in the counties they came from (although surprising little in my actual tree).

This means I often find multiple unexpected relationships to any given match. Turns out this is the norm for southerners and I've actually got it easy compared to most others who have shared their results with me!

This situation does NOT do well with a chromosome-focused approach.

Getting to that bottom layer of knowing where a segment came from is even harder than for "normal" people (I doubt there is a "normal" but the way genetic genealogy is taught, "normal" is people who's branches are distinct, not intermarrying in unexpected ways or more than once).

When you have more intermarriages in the population(s) in the test taker's tree, even confirming which branch a segment belongs to can be tricky. The larger the segment and the closer relatives you have as matches, the easier it is.

In this situation you can still rely heavily of identifying the correct branch and then bring in other types of confirmation. Remember, this isn't the same as identifying the shared ancestor based on the segment the ancestor came from (that's ancestor vs. branch if you aren't paying attention).

The problem is, you will have lots of times where you only have a family tree for a more distant relative and a smaller segment. Confirming the ancestor you both shared that segment from is much harder.

If you use a chromosome-focused approach, you can triangulate matches, but that doesn't mean you can find a shared ancestor.

In essence, the chromosome-focused approach jumps to the bottom layer and only supports it with the other layers as needed.

This can work but I've found it slow and boring due to the lack of confidence I can have when I find one shared ancestor for a small segment. Too often I end up deciding I've been chasing the wrong branch and am back at square one.

If you don't focus on segments, what do you focus on?

I hadn't considered a term for what I do instead of a chromosome- or segment-focus. The fastest explanation is to say I use a cluster-focus. This is because lots of people are familiar with auto-cluster tools. I don't rely on auto-cluster tools because of the same intermarriage-issue but my 4 Buckets Technique is a type of clustering and I've found it wildly helpful (otherwise I wouldn't waste my time doing it!).

Here's how I look at my layers.

I put LOTS of efforts into the top layer of building trees but also a very basic consideration of whether the shared cMs make sense with any identified relationship.

ALSO, I look at "data" in general (as opposed to for one test). This last item means I look at the shared cMs for multiple matches with the same shared ancestor and look for a pattern.

I have a statistics background but I don't do any serious data analysis when I'm looking for patterns. I personally consider that a waste of time if you aren't dealing with hundreds of data points, meaning hundreds of matches sharing that same ancestor. I don't have hundreds of matches to most identified shared ancestors although I do have hundreds I believe belong to a branch.

But believing a match belongs to a branch doesn't mean they don't also have DNA from an additional ancestor OR that they aren't related to an additional ancestor but don't share DNA from him/her.

There are lots of possibilities!

This emphasis on that top layer is exactly what I recommend in The Road to DNA Success. I recommend this regardless of where you tested.

What you do next depends on where you tested and what kind of matches you get.

I can only tell you what I do. I don't have a ton of experience outside of southern DNA. Even my husband's DNA is half southern and the other half has so many recent immigrants, I tend to get too few DNA matches (for any given branch) to willingly spend my limited time there instead of on my own family (hey, we all have to decide how to spend our time, genealogy isn't required so spend it how you enjoy).

I have found I get the most results working with AncestryDNA. That is due to the number of trees and having more, closer matches.

So, now we're back to GMP. Since it is chromosome-focused and AncestryDNA doesn't provide chromosome data of any type, it was very weak in what you could do with AncestryDNA results. I was mainly using it to try and connect (and keep organized) AncestryDNA matches that had results at other companies.

I was trying to take what I learned at AncestryDNA from trees and apply it to matches at other sites and get some insight using segments.

I'm not trying to go for "confirmation" based on the segment but add it as an additional layer. I'm more interested in identifying a branch based on shared segments (vs. identifying an exact ancestor). I'm also happy to exclude options (such as recognizing that a match with a known ancestor most likely has DNA from more than one shared ancestor and a group of matches, based on segment data, probably comes from the unidentified ancestor, not the identified one).

GMP for Ancestry DNA Analysis

So let's recap so you understand why I find this new GMP feature a "game-changer."

  • GMP has a chromosome-focus. Its features are aimed at identifying who a segment was inherited from and working from there.
  • With my ancestry, I find this difficult and dangerous because there is so often more than one (unexpected) relationship.
  • A chromosome-focus also doesn't work for AncestryDNA results since there is no chromosome data provided.
  • Likely you find the most trees (or info to start building a tree) at AncestryDNA which can allow you to know the most about AncestryDNA matches (I do).
  • Previously, GMP really just let you work with AncestryDNA matches you could identify at other testing companies. Other types of analysis really needed to be done outside GMP.

I've already told you the new feature allows you to see the shared match list (of AncestryDNA matches) for more than two people (for more than the test-taker and one match as you can do at AncestryDNA).

This feature essentially lets me do what I'm visually doing with my 4 Buckets Technique. I hadn't even considered trying to find a way to do this within Excel (yet).

The GMP feature basically let's me jump ahead and go more in-depth with the groups I've identified at AncestryDNA. Here's how it works for me.

I've focused on building trees for four years, at least. My 4 Buckets Technique uses the known shared ancestors as the starting place for "bucketing" (like auto-clustering but based on known shared ancestors AND matches can be in more than one bucket---I don't use a visual cluster to see the buckets or instances of multiple buckets which I find works better with results with lots of unexpected relationships although auto-clustering tools are addressing this more and more--- 'cause it's a serious issue).

I use Excel to "bucket" results quickly but I've found if I then use Ancestry DNA's colored dots to mark the buckets, I can actually assign new matches to the appropriate bucket(s) without leaving AncestryDNA (no need for Excel or any spreadsheets).

I use a method for assigning the limited number of colored dots that makes my buckets match the great-grandparents. The whole concept I started with was tossing matches into casual "buckets" to help me focus (initially four buckets, one for each grand-parent).

Early on I knew I could "sub-bucket" and that's what the eight great-grandparent buckets are. But I can sometimes go farther.

The new feature at GMP is exactly how identifying sub-buckets (more distant shared ancestors) works. You'll see people talking about sub-clusters or super-clusters with auto-clustering (super-clusters are where you put the smaller clusters back together which makes them more like my buckets, i.e. from a branch starting at a more recent point in your tree than the clusters start).

At least with southern DNA, some sub-buckets are easy to see (literally see) and some are quite convoluted.

This is to be expected when there are multiple shared ancestors. Keep in mind, sometimes the matches share multiple ancestors but don't share them with the test-taker. This adds an extra problem when you can't find the shared ancestor in the test-taker's tree but do find a shared ancestor between matches. That's why I focus on the shared cMs analysis and patterns as well as the buckets. They often clue you in to a problem even though you can't identify the exact situation.

The ability to generate a list of shared matches between three or more matches allows for sub-bucketing. This can identify groups who share the same more-distant ancestor in a given branch.

Knowing which branch to look in is a huge time saver!

Sub-buckets can also either highlight groups that share an additional ancestor or help throw-out matches that are false leads. The ability to create distinct lists and notice matches that appear (unexpectedly) on several, can be a major help.

You still have to do your own analysis but this additional level of AncestryDNA analysis is very helpful.

Remember, although you want to find new ancestors, saving time NOT chasing a false lead or just working on the wrong branch of the family is a very important result, too. With genetic genealogy we can easily be talking hours of time that is actually weeks or months you wasted (if you do "full-time" DNA analysis, your hours might span a few days, most hobbyists will be giving up a month or more of their limited DNA analysis time although the actual hours are the same).

Using GMP's AncestryDNA Analysis for "Focus"

With all DNA work, one of the issues is dealing with so much data. I've been struggling with the fact that I have helpful data in AncestryDNA (the "groups" represented by the colored dots). At this time, I don't have a way to download that information (if you know of a way, please let me know, this is a REALLY big deal to me).

 That means I have to manually enter these groups into GMP.

This new feature gives me more information but also lets me focus on a much smaller set of data to enter groups into.

With my great-aunt's test, her maternal branch has one lineage with every wife being a brick wall. DNA helped me identify an unknown line for the first brick wall wife (i.e. it identified the family of her mother).

 I've now got a lead on a second and possibly third new family (confirmed as belonging to the maternal side but not exactly which branch after that).

I've realized the only way to make more progress is to focus on one project. There are two ways to do this. One is focusing on an identified new branch (or a branch you want to make progress on).

With the 4 Buckets Technique, I can also focus on an adjacent branch and grow a sub-bucket to help exclude those matches from the initial project. Since I'm interested in all these unknown branches in the maternal side, I'm happy to do either. Sometimes when you get stuck it's necessary to do the latter (work on an adjacent branch).

GMP can now help me sub-bucket faster and more efficiently. The whole idea with the 4 Buckets Technique is you catch cases where there are multiple shared ancestors so you don't waste time chasing the wrong lead.

 It is possible my great-aunt's maternal side is even more inter-related than her paternal side (which I know has a lot of intermarriage among matches in addition to the 2nd cousins that married in our tree).

This new feature in GMP makes it much easier to investigate this and hopefully determine what is going on. I haven't had enough time to use it to say whether I'll make an actual breakthrough (without employing lots of other tools or finding a paper-trail answer).

This new feature is similar to how auto-cluster tools work. However, it allows you to control what you're looking at. It also won't fall pray to problems due to multiple shared relationship (your analysis can still run into the pitfall but I like this starting point much better).

For southerners, this is a huge deal, a game-changer.

In a nutshell

You can now use GMP to work with your AncestryDNA data in a similar way I do the 4 Buckets Technique.

  • You can do analysis not possible at AncestryDNA.
  • You can focus on a project instead of being inundated with too much data (as auto-cluster tools do).
  • You can work with sub-buckets (or sub-clusters) without causing matches to be excluded from additional sub-buckets they belong to (which is what happens with graphical auto-clustering).

Get Started

If you're concerned about multiple shared ancestors or have been overwhelmed using auto-clustering but like the concept, here's how to use this new GMP feature to get started with your AncestryDNA results.

  • Pick a project to focus on, preferably one where you've already identified a number of AncestryDNA matches as belonging to that project. Remember, you can work on an adjacent branch (or when you get started, any branch) to help exclude matches from other sub-buckets. You want to start where you have information.
  • Get your AncestryDNA data into GMP. Make sure it is up to date with matches you need for your project (it does not have to be up to date if you don't have any new matches for the project of interest but you will want to update it at some point).
  • Use the new analysis tool (explained in Becky's blog post about the new GMP update).
  • Make sure you first use the "Show Ancestry ICW" before the "Mutually ICW" option (the first is under the "Display" menu, the second under "Analysis Tools").

First, display your Ancestry shared matches
  • How many relatives you add to the Mutually ICW option is up to you.
  • Use your notes or "Family Group" to tell what's going on.
  • Update the notes/Family Group to get better and better results.
You can perform this on the Relative List or the DNA Comparison page. The DNA Comparison page is chromosome-focused and gives you a smaller display area so I'll stick with the Relative List, for now.

One of my "complaints" about GMP is the set-up of the Family Group field. This is where I get the most help for my AncestryDNA matches but it currently requires manual entry.

I'd like this to eventually have the option to update from my Ancestry groups (this first requires that info to be downloadable in a format GMP can accept which currently isn't even an option).

Most likely this would require a custom import template since not everyone uses the groups the same way but I know I could easily set that up for my own system, assuming I had the correct options to upload the info into GMP (i.e. I can set-up a csv to fit what I'd want to upload, but I can't alter what GMP will accept/which fields I'm allowed to import into).

Manually updating the Family Group is one advantage of working with a small number of matches identified in this process. I hate trying to do this for a ton of matches but when I'm going to get a use out of it right now, I'm much happier doing manual data entry.

Remember, you have a link in GMP to the Ancestry match page for each match if you need to refer to AncestryDNA's info (such as seeing what groups you've assigned to that match).

So there's (a lot more than I thought it'd be) about GMP and it's new Ancestry DNA analysis feature.

Let me know if you have questions. This is so new it's hard for me to give good examples. I'm used to working with my groups so I can see how helpful this feature is but talking about it generically isn't easy, yet. You can leave a comment or use the "Contact Me" form.

Now I'm going to get back to working with this new feature!

Get access to the free genealogy resource library by clicking here

No comments

Post a Comment