RootsChat.Com

Possibilities:

1) Incorrect ethnicity given
2) NPE with an Indian ancestor somewhere
3) Indian ancestry one generation prior, two at a push

Are there any ancestors who lived in India? It could be back another generation than the 1820s, but going back further is unlikely to turn anything more up.

Do they have any DNA matches who have ancestry from India, where those matches are also shared among themselves in a cluster? So you'd have, say 30 matches with ancestors in India, and they match with, on at least three in that cluster. Then look through the trees (expand incomplete trees as needed) and see if you can find MRCAs.

Obviously, there's not masses of Indians tested, and the ones that are tested are going to be from subsets of the population, e.g. Catholics, Brahmin, and Anglo-Indians.

But if there are no matches, it might be an error with ethnicity. You'd probably want to trace back another generation too.

I have tests with Indian ethnicity, and combining them it's difficult to see where it's come from. So I think their prediction with India is not as good as with British. There's also other groups in India that could theoretically get mix up as Indian. That happened with mine and FamilyTreeDNA. Armenian and Parsi are two that could get mixed up.

Hello,

I am looking for an extra pair of eyes on the cause of death on the below death certificate. The best I can get is - fistula.

Thanks for your time in advance.

Link: https://i.ibb.co/BWyBF83/TUDOR-Thomas-1842.jpg

I'm also getting the 8-19cM shared matches without having ProTools. This previously required a ProTools subscription. It seems odd that Ancestry are not making us pay for this. Christmas spirit?

Are you investigating the trees of your matches? Or are you mostly grouping?

If the group has a good amount of matches, it's often not difficult to find common ancestors among the matches. I open up dead-end ancestors of people in each tree and press the [Search] button for them. Typically they occur in another tree with more ancestors. Otherwise you can look them up in records. So the unknown ones I will label based on the common ancestors in the trees, such as:

Jones of Llanbrynmair
Dodd-Carter of Suffolk

With all the groups I don't know what the link is, I then put them into groups, where I think they probably have a connection, like so:



When you have groups of matches when the largest match is not greater than about 32cM (not sure of the exact threshold, might be higher), those tend to be more distant with common ancestors more in the window 1700-1730. That takes you back to about 6X GGPs, of which you have 128. So you'd need to have a pretty good idea of who those 128 are to make sense of your DNA matches, and I don't think many of us have close to that.

So, I've not been able to determine a relationship between most of my small clusters, which is in part due to remaining NPE gaps; otherwise due to inadequate records.

A key aspect is how far back shared ancestors in a cluster are. In my experience they always range from MRCAs in the window 1800-1660; typically more towards the former. If a cluster has matches that go back further on a line, then you can see there are no NPEs on that line.

You can have other clusters where you can only find MRCAs going back to 1800, and not know how they connect to you. So the common ancestor with you may be three or even four generations earlier. So when you don't know all your 6X GGPs and you then have to have a reliable tree for the 1700s of multiple lines of ancestry - most of which are not shared, it becomes cumbersome. If you are lucky there is an obvious surname link, but that occurs in the minority of cases.

As an example, see the char above where I have clusters for York and Ward-Wright. I found marriages between Yorks and Wards; and Wrights. However, since I don't know where in my tree that fits in, it's just a mystery.

There's also the possibility NPEs mean that actual paper trial in the cluster are false. There are so many variable that make DNA a delicate art. One ancestor was married in India and listed as being from an orphanage. Her surname was Nash and I've found nothing from DNA that connects to that surname. So I lean to that not being her birth surname. I've got a few clusters I know fit onto her general line, but I don't know which, if any, relate to her. One has MRCAs from about 1770. With no further generations. For all I know one of them was actually an NPE deposit from an unknown ancestor.

When I first began, with no knowledge, I clustered too fervently and made clusters that don't actually exist. This is done by clustering when there are not enough common matches. I tend to only put someone in a cluster if they share at least three matches in a cluster. If you add everyone who just matches one, you will end up with giant clusters with no common thread.

Another mistake I made is thinking that if I could not find DNA matches on a line, then it's probably due to an NPE. I felt pretty certain that two of my lines were NPEs. However, after more time and knowledge, I found that matches were just hidden away. Specifically almost all of the matches on one line also matched another line independently. Then the other cluster is just small and very diffuse; while still having matches with MRCAs back to c 1720.

So there's a very large amount of time and moderate knowledge that allows you to make the most of the matches. I was going through some matches a few days ago when after a lot of digging I found MRCAs from c 1715, some of which were from matches I'd written off. So there is always the possibility of finding more, but mostly I don't know how I am related to these people.

Sorry if these are basic questions. I have not been working with DNA result for very long.

1) On Ancestry (with ProTools), in the same way as Ancestry advises whether your matches are on the Maternal or Paternal side, is there any way of seeing whether a shared match is on the mat or pat side of the sharer?

They should really be providing this data as part of ProTools - what side of the match's tree do you match to. So rather than two full trees, you can narrow the search by 75%. That would require the match to have defined which of their sides is maternal/paternal.

The only way to get this information is to message the matches. If they are on a desktop, it will say if you are a maternal or paternal match in the message pane.