Ancestry DNA chromosome painter

[peter]

I notice there is an August update coming - the message I see refers to 8 new regions, though as I currently only have 4 out of 77 I'm not sure it will make any difference.

 

[Bob Spiers]

Fortunately there is a way to separate them: click on the names in the legend panel on the left and the colours will be turned on and off (greyed) accordingly.

Thank you Stuart that was a useful tip and makes a deal of difference. Much appreciated. And for the record your comment:

I think some here are believing too much of Ancestry's claims for this whole ethnicity thing, which is based on rather questionable assumptions,

I have always regarded Ancestry's Ethnicity proclamations as best guesses based (as you say) on questionable assumptions, and not to be taken too seriously. However I know from showing Ethnicity results to others whose trees I manage, they think them the 'bees knees' and they love the Nordic fringe ethnicities.

 

[peter]

this whole ethnicity thing, which is based on rather questionable assumptions

Which are?

 

[jorghes]

I don't really know how to use the Chromosome painter, but I, due to my (now well known) Jewish heritage, I can easily identify which parents are which in regards to my paternal side, for my results, my father's and my grandmother's - particularly as the Jewish only comes from one of my grandmother's grandparents.

So I'm happy to show you how any of it works if you want to know something. Just let me know what you want.

On the flip side, my mother's parents are virtually impossible to identify which is which, as they were both had an English parent, and a Welsh parent.

My latest results keep adding in strange additions - Norway, Italy and Sweden, none of which I have any evidence of anywhere in my tree.

 

[peter]

I'm happy to show you how any of it works

The question is whether it actually works.

For example, my wife's father was Welsh going back as far as we can get, but her mother had English grandparents on one side, and Welsh on the other. Comparing my wife's chromosomes with those of her mother I suspect that they have got at least one chromosome pair round the wrong way round the wrong way.

Is what you see consistent?

 

[jorghes]

The question is whether it actually works.

For example, my wife's father was Welsh going back as tar as we can get, but her mother had English grandparents on one side, and Welsh on the other. Comparing my wife's chromosomes with those of her mother I suspect that they have got at least one chromosome pair round the wrong way round the wrong way.

Is what you see consistent?

Well, all my great-grandfather's Scottish is on one side, and all my great-grandmother's English/Jewish on the other. They are all on the same sides of the chromosome painter, i.e. whichever (top or bottom) they have assigned to which parent.

I honestly am not completely sure what I am looking for, but I can always take some screenshots?

Attempt 2:
I have had a look at the "Painter" for my grandmother's results and my father's - particularly where the Jewish appears (the Scottish is pointless for my father, his father also has Scottish blood!). My father's Jewishness only appears on Chromosomes where his mother's results has Jewishness on the same Chromosome.

Is that the sort of thing?

 

[peter]

Ancestry shows that my wife has inherited English DNA from both parent 1 (chromosome 14) and parent 2 (chromosomes 10 and 22). As

 

[James738]

I'm manager for mine and my 2 siblings DNA.
In general, our results are boringly English and Scottish (90-95%). ONE single split origin chromosome between us, and its not even mine .
We knew that my mum had one Scottish grandfather with all the others from Middle England. My dad's family as far back as the 1700s were Huntingdonshire and Cambridgeshire.
Yet I and my brother have Scottish ethnicity on both sides. Mine just for chromosome 7, his for 7, 17, 19 and 20.
My sister only has Scottish ethnicity on the lower chromosome in her image (parent 2).
Ours are primarily the upper gene (parent 1). It turns out that Ancestry randomly allocates parent 1 to one side or the other. Whilst you can edit parent 1 and 2 to show them instead as paternal or maternal, when switching views between siblings, they remain in the original randomly set order. I've suggested that it would be useful to allow us to sort them the same.
Ethnicity estimates are definitely a work in progress. Both siblings are showing a full paternal chromosome as Welsh (13 and 18), my brother also one maternal (8). Their Welsh inheritance is actually only 2%, i.e. a range of 0% to 9%, so perhaps it doesn't exist. Yet in each case the full chromosome is coloured for Welsh. If anything I'd have expected maternal because on her side a 3rd ggf came from Gloucestershire.

 

[Bob Spiers]

Which are?

Although addressed to Stuart, as I agreed with his 'questionable presumptions' when responding to him, here is what I meant.

I regard all statistical predictions as best guesses at likely outcomes -the weather forecast -global warming-performance of the economy- Covid predictions etc., etc., - and the more information the statisticians have to hand the better the guess. Ancestry is head and shoulders above their rivals, so one must accept their 'guestimates' are likely (but not certainly) better than many others working on less information.

In the same way 'Straw Polls' taken at Polling Booths on Polling day are much more likely to predict an accurate outcome, than (say) Newspaper Polls taken a week or two before. But the true result and true science can only come after the voting slips are counted and the overall result given.

That is why it is harsh, but fair, to regard Ancestry predictions (or presumptions) as 'questionable' because we -and certainly Ancestry in explaining how 'Side View' works: "We can split up your DNA according to the two parents who passed it down, but we still can't connect each half to a specific parent" - know the limitations of what they offer...so must we.

 

[peter]

here is what I meant

Bob, I think it's important to differentiate between predictions, which are about the future, and Ancestry's ethnicity estimates, which are very definitely about the past.

It's also worth noting that even if Ancestry's algorithm was 100% accurate they would be unlikely to accurately identify which parent was which, because the best way of achieving this is by reference to epigenetic markers, information that is not captured using low-cost genealogical DNA tests. Accordingly their statement has no bearing on the accuracy or otherwise of their algorithm.

In many, probably most, cases it's obvious which is which

 

[Singer35]

I'm here through Lost Cousins newsletter. I am blessed that my ethnicity is clearly defined for both my deceased parents and they are not interconnected at all. My maternal DNA is my Liverpool born mother and paternal is my French-Canadian father so thankfully I was able to recognize Maternal and Paternal in the early version when they were named Parent 1 and Parent 2. They have some small overlaps in Sweden-Denmark and Scotland but on different chromosomes and probably so far back I'd never trace them genealogically. I am wondering if anyone has experience with Gedmatch and how I might use my mother's chromosome data to the best advantage to break down my English brick wall? Will post my brick wall elsewhere.

 

[JFB]

I am underwhelmed by the "new" Ancestry DNA painter.
Why can't they just provide a proper tool like other sites do instead of a dumbed down version for marketing purposes. They have the necessary data, they have more of it than any other site, so why do I have to export my DNA data to other sites to use it properly - at least Ancestry still allow that (for now).
I am becoming increasingly disillusioned with the lack of development of useful tools by Ancestry. Most of their updates so far this year have been presentational and "pretty". They are falling behind their competitors, rapidly.
Their new chromosome painter tells me nothing useful, both parents 95% from northwest Europe. Whereas if I could paint DNA strings from my shared matches together myself that would be fantastic. For a lot of matches I know where they connect in my tree, or which "cluster" they belong to, so it would help me to pinpoint and confirm likely common ancestors for the unknowns. I can do this at other sites, but their data bases are somewhat more limited.

 

[KirstenB]

I too have just looked at the ethnicity painter as a result of Peter’s newsletter. Before I look at the full analysis of the 5 tests, I manage I too wondered whether it would help with Gedmatch. If, as seems the case, Ancestry is suggesting a specific ethnicity on a specific chromosome AND roughly its position on that Chromosome, it would be interesting to check which Gedmatch matches I have in those places. E.g., are people I have identified on Gedmatch as relating to my only Welsh line, showing up on Gedmatch on Chromosomes, where Ancestry have suggested I have DNA similar to their Welsh community group. If that does work, then maybe possibilities (rather than certainties) could be suggested by other section matches. Hypothetically (as I don’t have a wide-ranging ethnicity profile), if I had a group of matches on Gedmatch in a particular section of a say Chromosome 2, and I hadn’t identified where they belong in my tree, if Ancestry suggested that this part of Chromosome 2 area matched a Spanish community, maybe those Gedmatch matches are on my “Spanish side”?

In relation to Ancestry’s example, I guess if it is genuine, it may be a USA use citizen. I suspect as many USA citizens may have a more varied mix than many of us Brits, having the DNA painter may be more help to them, who are, after all, Ancestry’s main client base. Also, maybe if USA citizens have more varied result the colours on their chart are easier to distinguish. Not much help to us with all pale green and blue variations. You could copy the chart as an image and change the colours.

 

[KirstenB]

I also see that by clicking on the colour in the chromosome or in the list (for, all, parent 1 or parent 2) it highlights that ethnic grouping (it lets you do one or more.) In the case of looking at parent 1 or parent 2 it will highlight in a paler version where the other parent has the same ethnic group on the chromosome)

 

[peter]

They are falling behind their competitors, rapidly.

Ancestry are streets ahead of their competitors, and I don't see them losing their lead. I judge the different sites by the results they produce, not by the tools that they offer.

 

[kctoo]

Hi Peter et al, I have access and additionally also tested both parents.
I can report that, like ethnicity inheritance before it, there are some differences between Ancestry's prediction and the actual result.
For background, all my 3 x great grandparents left the Uk, with the exception of 2 from Ireland, in the 1800s for Australia/New Zealand and according to Wikitree my parents are 14th cousins.
Ancestry's prediction for me, for parent 1 and parent 2 included DNA for Wales on Chromosome 11 and 20 and on Chromosome 8 for parent 2. However when I checked my parents "actual" Chromosome Painter, my Dad had no DNA markers for Wales and my mum had markers on no.s 4, 8 and 12.
For my DNA from Sweden and Denmark, I have some predicted for chromosome 17 on parent 2 but none appears for either.
With my Norwegian DNA there was more success. Ancestry predicted parent 2 would have some on no.s 4 and 12 and my mother indeed had exactly that with the addition of some on 15.
As for my Scottish DNA, the prediction was for my Dad to only have 1 chromosome however his own results show another 10.
I have of course yet to report the results of the English and Irish DNA but I think you get the gist.
Hope this is of some interest.

 

[peter]

Thanks for your input, kctoo.

From what you've told us it sounds as if, in your case, parent 2 is your mother. This is supported by the Norwegian ancestry on chromosomes 4 and 12, and the Welsh ancestry on chromosome 8.

It's important to remember that we only inherit half of our parents' DNA, so the fact that your mother's chromosomes 4 and 12 also show Welsh ancestry is perfectly consistent. What isn't consistent is the Welsh ancestry indicated by your chromosomes 11 and 20.

It would be interesting to know whether what you can see for your parents fits with what you know about their ancestors.

Incidentally you can safely ignore what Wikitree tells you about your parents being 14th cousins - even if it is correct, which is unlikely, they will be cousins many times over, and some of those cousin relationships will probably be closer than 14th cousin.

 

[KirstenB]

kctoo's results are interesting and the type of inconstancy Peter referred to I have also found in my results for one Chromosome. I have tested and so has my mother. I am fairly certain my father is my parent 2 as that has strong Welsh ethnicity results compared to parent 1; my paternal grandmother was Welsh. That leaves my mother as parent 1. On Chromosome 19 for parent 1 Ancestry suggests Scottish ancestry but my mother has English predicted for both her parents on Chromosome 19.

I haven't found any Scottish ancestry on either side of my family. There does seem to be this Ancestry tendency to find Scottish ethnicity in many results -(overall it gives me 10% Scottish). The majority of both sides of my family (7 of 8 great grandparents and before, back to between 2x ggps and 6x ggps) come from roughly NW England/NE Wales and surroundings (South Lancs, Flintshire, Cheshire and Derbyshire.). The other line, for my paternal great grandfather (identified back at least geographically beyond 10xggps) is more SW/South Central England (mainly Wiltshire, Somerset, Hampshire-some Berkshire).
Overall I think the new DNA painter is interesting but, as you would expect not completely accurate.

 

[peter]

There does seem to be this Ancestry tendency to find Scottish ethnicity in many results -(overall it gives me 10% Scottish)

They do explain this, and your Scottish component is actually below average for someone of British origin..

 

[kctoo]

It all fits Peter, except for the Welsh ancestry. I didn't know I had any specifically. But then I did have a couple of brick walls with the name of Williams on mums side. One in Birmingham and one in Bristol. Benjamin Hook and Elizabeth Williams who married at St Nicholas Bristol 4 Mar 1777 are the most likely parents of four children baptised at St John the Baptist Bristol including my 4x g grandfather Edward bap 21 Feb 1788. His twin brothers bap 14 Jul 1782 were recorded as being approx 6mths old at the time. Unfortunately with that marriage is the end of that line for me as I can find no likely baptism records for Benjamin in a 20 mile radius and with Elizabeth ironically it is the reverse. As you would expect there are far too many possible Elizabeth Williams baptised in the area for me to even hazard a guess at which might be mine. A few years later and she might have survived to see the '51 and I might have had some luck, but no dice. Not an occupation or birthplace in sight for either of them. But perhaps now DNA has given me clue as to where she or her parents or grandparents are from, not that any brick walls will be broken with it this time round.