SideView

[peter]

So far, every time I select "by parent" I just get "Something went wrong. Please try again later.".

Exactly - that's why we can't see the explanatory information. If I'd known this was going to happen I would have taken a screenshot.

 

[Pauline]

So far, every time I select "by parent" I just get "Something went wrong. Please try again later.".

Same happens to most people, it seems, but if you click the dropdown on Groups you get the option there to select Parent 1 or Parent 2, and that works - most of the time, anyway!

 

[Pauline]

Presumably that's a match where you had assigned it as paternal or maternal?

Yes, based on the relationship I either knew I had, or had worked out, with that match. But because I had done this previously, I now know which parent is 1 and which is 2, something I couldn't tell just by looking at the ethnicity estimates.

It may also be a useful pointer with more distant matches, as although it may be clear from trees that we are related, the DNA we share may not necessarily come from that relationship.

And it's useful with unknown matches where I have shared matches on both sides, as it clarifies where I should be looking for our relationship. For example, there is a match in my sister's list with whom she shares around 120 cM, and with whom there are shared matches with documented cousins on both my mother's side and my father's side. This new feature tells me we share DNA on my father's side.

 

[trebor]

I think that this will become very useful when the bugs are ironed out but at the moment I am somewhat confused as it is saying that my maternal 1st cousin is "both sides".
So far I have found no crossover families that could indicate this and I have gone way back on most lines. My maternal side is Staffordshire and surrounding area while my paternal side is Co Durham / N Yorkshire which, to me, looks like the two were not likely to meet. Admittedly there are a few "unknown" fathers on each side but for this to be the answer would be a big coincidence.

If anyone can explain this to me I would be pleased to hear the reasoning.

 

[peter]

I think that this will become very useful when the bugs are ironed out but at the moment I am somewhat confused as it is saying that my maternal 1st cousin is "both sides".
So far I have found no crossover families that could indicate this and I have gone way back on most lines. My maternal side is Staffordshire and surrounding area while my paternal side is Co Durham / N Yorkshire which, to me, looks like the two were not likely to meet. Admittedly there are a few "unknown" fathers on each side but for this to be the answer would be a big coincidence.

If anyone can explain this to me I would be pleased to hear the reasoning.

If I remember rightly from when the explanatory information was still available (ie yesterday evening) a match is shown as both sides if there are at least two DNA segments from each side. Since you're likely to share 35-40 segments with a 1st cousin the 95% accuracy claimed for SideView comes into play - 5% of 40 is 2.

 

[Helen7]

Like Pauline, I don't yet have this feature in my account. However, if I log in to my husband's account, I can see it, and I can view both our tests. In both cases, those marked Parent 1 match those I'd assigned to the paternal side and those marked Parent 2 match those I'd assigned to the maternal side, which is reassuring. It also gives me some confidence in assigning as maternal/paternal those I'd previously no idea which side they were.

Surprisingly, my husband has an approximately equal number assigned to each side and unassigned (each just over 6,000) whereas I have over twice as many on my paternal side (11,000) than on maternal side (4,700) with 7,500 unassigned. The only one assigned to 'both sides' is our son, as expected.

I need to investigate further, but looks to be a useful tool.

 

[trebor]

If I remember rightly from when the explanatory information was still available (ie yesterday evening) a match is shown as both sides if there are at least two DNA segments from each side. Since you're likely to share 35-40 segments with a 1st cousin the 95% accuracy claimed for SideView comes into play - 5% of 40 is 2.

Thanks Peter but still not making much sense to me. Are you saying that this match comes up in the 5% inaccurate section?
If that is the case I would be cautious about the much smaller matches which make up the majority.
If it helps, my match with my cousin says 15% shared: 1033 cM across 33 segments.

 

[peter]

If it helps, my match with my cousin says 15% shared: 1033 cM across 33 segments.

5% of 33 is 1.65, which means it's quite feasible that 2 segments were allocated to the wrong parent.

 

[trebor]

5% of 33 is 1.65, which means it's quite feasible that 2 segments were allocated to the wrong parent.

Thanks again - I will keep an eye on this to see if the accuracy improves

 

[peter]

Thanks again - I will keep an eye on this to see if the accuracy improves

It'll be a long wait. This high level of accuracy is only achievable because they have over 23 million tests in their database - for a significant improvement in accuracy they'd probably have to sell at least 10 million more tests.

But why does it matter, since you know how you're related to that cousin anyway? Surely the value of SideView is its ability to identify how you're related to more distant cousins?

 

[Stuart]

Now I do have that "by parent" page working, so I have numbers - 6,006 for Mum, 10,188 for Dad, and 4,778 don't knows. That split susprises me a bit, but then it it mostly a count of very distant matches that I've never pinned down at all.

There is now a support page "DNA Matches Split Up By Parent", which suggests that the little (i) icon will stay after you've used it and the relabelling is reversible.

 

[canadianbeth]

When I click on the "Beta" tab it breaks down how many of each parent matches I have. For Parent 1 (maternal) I have 7092, Parent 2 (paternal) 6177, Unassigned 2766, and one for Both ( that is my niece - sister's daughter) Last evening when I first looked I had just 2757 Unassigned so between then and now nine more have been added, probably down in the 20% or less group.

 

[Helen7]

For example, my mother-in-law, who had one English parent, and one Welsh parent, has very few unassigned matches - about a tenth as many as most of the other tests I manage.

My husband had one English parent and one Welsh parent, but he still has 6,335 unassigned matches, more than either English Parent 1 (6,209) or Welsh Parent 2 (6,185), so 33% unassigned. Both my parents were English but I have a similar proportion (32%) of unassigned matches (7,497 unassigned, 15,835 assigned).

 

[Sue_3]

I think that this will become very useful when the bugs are ironed out but at the moment I am somewhat confused as it is saying that my maternal 1st cousin is "both sides".
So far I have found no crossover families that could indicate this and I have gone way back on most lines. My maternal side is Staffordshire and surrounding area while my paternal side is Co Durham / N Yorkshire which, to me, looks like the two were not likely to meet. Admittedly there are a few "unknown" fathers on each side but for this to be the answer would be a big coincidence.

If anyone can explain this to me I would be pleased to hear the reasoning.

I also have my maternal first cousin listed as 'both sides'. They are the only match of mine that is listed as 'both sides'. They are also my closest DNA match, which is unsurprising since neither of us has any known closer relatives still alive. This fits with the explanation offered by Ancestry, via Peter, that with a lot of shared DNA there is increased likelihood that some of it falls into the 5% of erroneous allocation. However, I also know from another site that both mine and my cousin's DNA has been uploaded to that my maternal cousin does indeed share a small amount of DNA with someone who is a match on my paternal side. I think that might possibly be because they both have a lot of Irish ancestry (although I don't - my maternal cousin gets theirs from their parent that is only related to me by marriage and I get the small amount I do have from my paternal side) so there's quite a lot of scope for them having some distant connection?

 

[peter]

the explanation offered by Ancestry, via Peter, that with a lot of shared DNA there is increased likelihood that some of it falls into the 5% of erroneous allocation

That's my interpretation of what I know about how Ancestry classify a cousin as unassigned - they wouldn't necessarily agree with what I've said.

However, I also know from another site that both mine and my cousin's DNA has been uploaded to that my maternal cousin does indeed share a small amount of DNA with someone who is a match on my paternal side. I think that might possibly be because they both have a lot of Irish ancestry (although I don't - my maternal cousin gets theirs from their parent that is only related to me by marriage and I get the small amount I do have from my paternal side) so there's quite a lot of scope for them having some distant connection?

Most distant connections involve a single shared segment of DNA, not two segments. So it's very likely that at least one of the segments has been wrongly allocated. Even if one of the segments has been inherited on your father's side, if it's a result of his Irish ancestry it's unlikely to be meaningful.

 

[peter]

CLARIFICATION: the information about how matches are designated didn't come directly from Ancestry, it comes from the blog post which I referenced earlier. It is a MUST READ.

 

[PhilGee]

My first 10 matches in the list are all paternal and range from 634cM to 91cM and 6 have a circled "i" - 2 out of 3 on my father's paternal side, including the 17 segment 634cM (my fathers first cousin) and 4 out of 7 on my father's maternal side (all 7 descended from the same sister of my paternal grandmother).

The first match without a "paper trail" link is 50cM and all shared matches are "father's side". I have just removed my "relationship" to see what Ancestry decides at the next update - it has changed to "Unassigned" for now.

I shall have to re-read the blog post now I have seen Ancestry's choices.

 

[jorghes]

So now the parent pages are working for me:

Test 1 (Mother - Welsh/English)
Parent 1: 10,219
Parent 2: 8,174
Unassigned: 3,995

Test 2: (Father - Scottish/English/Jewish)
Paternal (1): 8,271
Maternal (2): 9,324
Unassigned: 1,456

Test 3: (Paternal Grandmother - English/Scottish/Jewish)
Maternal (1): 29,161
Paternal (2): 16,599
Unassigned: 2,896

Looking forward to checking if I can properly assign each of my maternal sides for my mother's test. Heading off to explore the results now.

 

[peter]

I was intrigued to see this match in my mother-in-law's results:


Given that we know a 'Both sides' match requires at least 2 segments from each side it seems impossible. However it turns out that TIMBER had downgraded the segment length:

 

[Sue_3]

That's my interpretation of what I know about how Ancestry classify a cousin as unassigned - they wouldn't necessarily agree with what I've said.

OK, you referred to the explanatory information that was briefly available on Ancestry, which I haven't seen. I had read the blog. I couldn't tell how much of your post came from which source or was just your opinion, so I ought to have left Ancestry out of it! My cousin is in the 'both sides' group, not 'unassigned', but labelled as 'Maternal side' (without a little 'i') in my 'all matches' list.