Newsletter - 17th May 2018
Ancestry DNA offers END SUNDAY/MONDAY
The LostCousins newsletter is usually published 2 or 3 times a month. To access the previous newsletter (dated 11th May) click here; to find earlier articles use the customised Google search below (it searches ALL of the newsletters since February 2009, so you don't need to keep copies):
Whenever possible links are included to the websites or articles mentioned in the newsletter (they are highlighted in blue or purple and underlined, so you can't miss them). If one of the links doesn't work this normally indicates that you're using adblocking software - you need to make the LostCousins site an exception (or else use a different browser, such as Chrome).
To go to the main LostCousins website click the logo at the top of this newsletter. If you're not already a member, do join - it's FREE, and you'll get an email to alert you whenever there's a new edition of this newsletter available!
The world of genealogy has been rocked by the news that Ancestry, the biggest genealogy company the world has ever seen, is being sued by 23andMe, whose co-founder and CEO is Anne Wojcicki, the ex-wife of Google co-founder Sergey Brin. Entitled "Finding relatives in a database" the patent's first claim is for:
"A method for determining a relative relationship of people who share a common ancestor within a threshold number of generations, comprising: obtaining recombinable deoxyribonucleic acid (DNA) sequence information of a first user and recombinable sequence DNA information of a second user; wherein the recombinable DNA sequence information of the first user and the recombinable DNA sequence information of the second user are stored in a database comprising recombinable DNA sequence information of a plurality of users; determining, using one or more computer processors, based at least in part on a comparison of the recombinable DNA sequence information of the first user and the recombinable DNA sequence information of the second user, a predicted degree of relative relationship that corresponds to a number of generations within which the first user and the second user share a common ancestor; and notifying at least the first user about the relative relationship with the second user."
I couldnít help noticing that if you substitute census information for DNA sequence it's a pretty good approximation to what LostCousins does - and has been doing since 2004. You can read the entire patent if you follow this link.
I canít judge the merits of 23andMe's case, but I'm glad that I've already tested myself, my brother, and several of my cousins with Ancestry DNA. If 23andMe were to win then Ancestry would probably have to pay 23andMe a licence fee, and this would inevitably push up the price of their tests. But the more immediate impact could be the further delaying of Ancestry's plans for an IPO, which have already been postponed once.
Coincidentally I recently tested with 23andMe after buying a kit during their DNA Day sale - it will be interesting to see what I get for my money. Somewhat confusingly the basic 23andMe test, the one I bought, is called an Ancestry test (to distinguish it from their more expensive Health and Ancestry offering).
Genealogy sites close down thanks to GDPR
This week I received an email from the Family Historian User Group to let me know that they're discontinuing their newsletter as a result of the new data protection regime - but at least the FHUG is continuing to support users.
On the other side of the Atlantic several sites which support genealogists are closing down ahead of GDPR, reportedly including Ysearch.org and Mitosearch.org (according to the ISOGG wiki and other online sources, though there was no mention of this on the sites concerned when I checked this morning). Also closing down is another DNA-related site, World Families Network - I've never used it myself, but apparently it hosts a number of DNA projects.
In the case of Ysearch and Mitosearch it could be said that they've outlasted their usefulness - both were set up by Family Tree DNA to allow people who had tested their Y-DNA or mtDNA with other providers to match with those who used FTDNA, but now FTDNA is effectively the only source of those tests. You can read more about these closures in Roberta Estes' DNA Explained blog.
In the last issue I mentioned that I had been recently selected for jury service, and as a result some useful tips were sent in by readers who had previous experience, some in the UK and some overseas.
The concept of jury trials was first introduced into English law by clause 39 of the Magna Carta of June 1215, but it was actually a decision by Pope Innocent III in November 1215 which put the final nail in the coffin of the mediaeval 'trial by ordeal' system, when he forbade clerical participation - there is more historical information in this article by Geoffrey Robertson QC at the British Library website.
But one interesting fact you won't find in that article is that if insufficient jurors turn up at the start of the case itís possible to select passers-by or people working in a local shop or office to fill the vacancies. This process, known as 'praying a tales' or 'praying a talesman' (sometimes incorrectly described on the Internet as 'paying a talesman'), was used as recently as 2 years ago, when 3 jurors from a panel of 14 at Salisbury Crown Court were unable to take part - one knew the defendant, one knew a detective involved in the case, and one had fallen ill (see this June 2016 article from the Telegraph). A number of law revision sites refer to a case at Middlesex Crown Court in 1992 when around half of the jurors failed to turn up after the New Year holiday, but I haven't been able to find a primary source to verify this.
Note: the practice of 'praying a talesman' to fill a vacancy carried over from English law to other English-speaking jurisdictions, including the US and Australia.
A much less onerous task is
to act as a witness to a marriage - this is another role that is sometimes
filled by a bystander. Register office staff are not allowed to be witnesses,
but otherwise anyone
aged 16 or over who is mentally-competent and understands
English can be a witness - illiteracy is not a bar.
Saturday is a big day in the UK - and I'm not just referring to my presentation at the Essex Society for Family History in the morning. Apparently a couple of 30-somethings called Harry & Meghan are tying the knot around mid-dayÖ.
To celebrate the Royal Wedding Ancestry.co.uk are discounting their tests to £63 (plus shipping) from Friday 18th May until Monday 21st May, and there are also offers in Australia and Canada - see below.
Please use the links below (and make sure tracking hasn't been disabled) so that LostCousins can benefit from your purchase - it wonít cost you a penny more, but it will help to keep this newsletter independent:
Ancestry.co.uk (UK only) £63 plus shipping ENDS MONDAY
Ancestry.ie (Ireland only) Ä76 plus shipping ENDS MONDAY
Ancestry.com.au (Australia and New Zealand only) AUD$99 plus shipping ENDS SUNDAY
Ancestry.ca (Canada only) CAN$99 plus shipping begins noon Friday, ENDS MONDAY
Note: I've updated this Masterclass since it was last published in February
Before you begin - forget those ethnicity estimates
Ethnicity estimates are for amusement only; even if they were correct, and itís very unlikely they are, the areas are often large and ill-defined, and the timescale is usually 1000-2000 years, well outside of the period for which records exist.
Introduction to autosomal DNA
If youíre male and you've previously tested your Y-DNA you might be expecting a set of numbers, For example, I tested 111 markers on my Y-DNA and I can see from my results how many repeats I have at each of those 111 sites - comparing my Y-DNA with that of other males who have tested is easy, and when I do the number of differences provides a rough guide to how closely related we are.
Autosomal DNA is very different - your DNA is sampled at around 650,000 to 700,000 sites, and there are two readings for each site (because the autosomal chromosomes come in pairs). Looking at the raw data isn't going to tell you anything - it takes a clever computer program to compare your results with those of millions of others who have tested, identifying common segments of DNA. But even when those segments have been identified, there's nothing to say which ancestral line they came from - there are no names attached, nor any dates.
So no matter how much experience you might have as a family historian, it would be understandable if, when the results of your DNA test came through, you were completely flummoxed about what to do next. There's a simple reason for this - we're used to working backwards from what we already know, so there's a clearly defined path, ie: find our ancestor's baptism in order to discover (or confirm) who their parents were, then find the parents' marriage, then find the baptisms of the parents and so on, working back a generation at a time.
But when we're matched with a genetic cousin, someone who appears to have inherited an identical segment of DNA, we're faced with a very different challenge - we usually donít know which of our ancestors we inherited that segment from, and the chances are the person we're matched with won't know either. It's rather like trying to do a complex jigsaw without first seeing the picture on the box.
Most of the matches we make with DNA cousins will be many generations back, since we have many more distant cousins than we do close cousins. The final column of the table below indicates roughly how many cousins you might expect to find if you and they all took the Ancestry DNA test:
Based on Table 2 from: Henn BM, Hon L, Macpherson JM, Eriksson N, Saxonov S, Pe'er I, et al. (2012) Cryptic Distant Relatives Are Common in Both Isolated and Cosmopolitan Genetic Samples. PLoS ONE 7(4): e34267. doi:10.1371/journal.pone.0034267
Revised using Ancestry DNA estimates for the chances of detecting cousins and the expected number of 1st to 6th cousins for those of British ancestry; the numbers for 7th to 10th cousins are my own guesstimates
Of course, in practice only a small fraction of your cousins will have tested - even Ancestry, by far the biggest providers of autosomal tests, have only sold about 10 million tests - but you can nevertheless reckon that the cousins you're matched with will be distributed roughly in proportion to the figures shown above. In other words, over 99% of your matches will be with relatives who are at best 5th cousins, and could well be 8th cousins or even more distant.
Tip: Ancestry won't show any of your DNA matches as more distant than '5th to 8th cousin', but it's very likely that amongst them there are many who are more distant - possibly up to half of them. Once you get beyond 3rd cousins the length of the shared segment(s) is only a very rough guide to how closely you are related - you could share a 7cM segment with a 10th cousin, but no detectable DNA with a 5th cousin.
This amazing chart from Blaine Bettinger's blog shows how variable the amounts can be, and how this affects the amount of DNA shared by more distant relatives:
In each box there are three figures: the lowest and highest amounts shared between relatives of each order, together with the average. However the average only takes into account matches - if there was no detectable shared DNA then that example was ignored.
What you will notice is that the average stabilises at around 12 or 13cM even for the most distant relationships in the chart. The average DNA shared between 8th cousins is just 0.055cM, but the average in the table is over 200 times greater. That's because unless there's a matching segment of at least 6 to 10cM most companies won't report a match at all - and this brings up the average, which only includes matches which were actually detected.
Very interesting, you might think - but what does it actually mean in practice? What it tells us is that neither you, nor I, nor any of the DNA companies can reliably predict how closely we are related to our more distant cousins.
Even if your DNA match is with a 5th cousin, someone who shares your great-great-great-great grandparents, it probably wonít be obvious how the two of you are related. I don't know about you, but I certainly can't say who all of my 4G grandparents were - indeed, I don't even know for sure who all my 3G grandparents were. I've got several 'brick walls' in the last 6 generations - and most researchers, including my DNA cousins, are probably in the same situation. Go back another generation and there are even more gaps - and it just gets worse from then on.
In other words, most of the ancestors that link us to our DNA cousins are on the other side of a 'brick wall' - and this could be a 'brick wall' in your own tree, in your cousin's tree, or even in both trees. What a challenge!
At this stage it's important to remind ourselves why we took a DNA test! Surely the primary reason we tested was to knock down 'brick walls' that conventional research couldn't breach? If our 'brick walls' have resisted our efforts for years (or even decades), the opportunity to knock them down using DNA is surely well worth grasping - even though it will mean that we have to adopt a new and unfamiliar strategy, and utilise somewhat different techniques?
Before you get your results
Make sure that you've done all the conventional, records-based, research you possibly can. Remember, DNA testing isn't a substitute for records-based research - you need to do both to have a reasonable chance of success. Each builds on the other - if you only do one you're almost certainly going to fail.
Complete your My Ancestors page, ensuring that you have entered ALL of the cousins (no matter how distant) that you can find on the 1881 Census. Yes, it might take you an hour or two, but skipping this important step could cost your tens or even hundreds of hours when you come to analyse your DNA matches.
Tip: start in 1841 and trace each of your branches (sometimes referred to as collateral lines) through to 1881.
Take a look at your My Cousins page and see which of your cousins have already tested, then contact them and find out who they tested with. If they tested with a different company, ask if they have uploaded their results to GEDmatch, and if so, what their kit number is. Shared matches are the easiest way to figure out how youíre related to DNA cousins, so knowing which cousins have already tested is crucially important.
How to process your DNA matches
I'm going to assume for the purpose of this article that you tested with Ancestry - but don't stop reading if you tested elsewhere because I'll be covering techniques you can use, though not as effectively, at Family Tree DNA and GEDmatch.
At Ancestry you'll typically have 10000 to 20000 matches with cousins, and of those all but about 1% will be with 'distant' cousins, ie where the estimated relationship is 5th cousin or more distant. So you might think that the best strategy might be to focus on the 1% on the basis that if you can't make head or tail of those matches, your chance of resolving the more distant matches is negligible. Wrong, totally wrong - that approach will lead to frustration, disappointment, and disillusionment!
Here's how to get the best results and avoid all that wasted time and frustration:
Strategy 1: search by surname
My experience has shown that a much better approach is to search the trees of your matches by surname, in the hope of identifying cousins who have the same surname in their tree as one of your 'brick wall' ancestors. Here's how to go about it:
Strategy 2: search by birthplace
As you will have discovered when working through your list of surnames, most of the time the surname of the ancestors you share with a DNA cousin doesn't appear in both trees - indeed, it's quite possible that the surname of your common ancestor doesn't appear in either tree!
The problem is, when your female ancestors married they generally took their husband's surname. This makes it more difficult to research female ancestors whose children were born before the commencement of civil registration, since baptism registers don't usually give the mother's maiden surname - usually the only solution is to find the marriage. By contrast you can continue researching your male ancestors even if you can't their marriage.
Of course, this problem doesn't simply affect you and your research - it affects your cousins too; most researchers' trees become increasingly sparse with each generation. If you've only identified 10% of your 256 6G grandparents and your cousins have only identified 10% of theirs, the odds of finding out how you're related to a 7th cousin simply by comparing the names in your trees are pretty remote (a little more than 1% in this example, not great odds).
Another way to figure out how you are related to your DNA cousins is to look for geographical overlaps - and here's how to go about it:
Strategy 3: look for overlaps with the
more unusual components of your ethnicity
Most readers of this newsletter have mostly British, Irish, or western European ancestry. But some of you will have Jewish ancestors, or ancestors from outside Europe, and whilst ethnicity estimates can be quite misleading, they do provide another way of analysing your matches.
Here's what Ancestry show for one of my DNA cousins:
If Ancestry had detected a Jewish component of my own ethnicity this would be one of the matches I'd be looking at very closely.
Strategy 4: look for the 'elephant in the room'
Because we all have 'brick walls' in our trees there are parts of our ancestry that are a closed book - yet there will inevitably be clues amongst our matches, if only we look for them. For example, if - like me - you don't know of any Irish ancestors, but have lots of matches with cousins who do, you might begin to wonder whether one of your 'brick walls' is concealing a connection to Ireland. I can't provide you with a step-by-step guide - it's all about awareness (Louis Pasteur said that "chance favours the prepared mind").
But beware of the common situation in which you share the same DNA segment with lots of other people (if you tested with Ancestry you won't know about this unless you upload your results to GEDmatch and use the Chromosome Browser). This can indicate that the people youíre matched with come from an endogamous population, one in which people generally marry within the same community - in this case you would probably do well to ignore the matches altogether as any connection is likely to be a long way back.
Testing your own DNA is a great start - but itís only the beginning. It's unlikely you'll be able to figure out how you connect to more than a fraction of 1% of your DNA cousins if all you've got to go on is a single test. How can you possibly make use of the remaining 99%?
The key problem, as the Masterclass points out, is not knowing which of your ancestral lines connects you to each of your genetic cousins. However, if one of your cousins has tested you'll know that the matches you share with them are on one of the lines that you share.
For example, if a maternal 1st cousin has tested you'll be able to identify some of your matches as being on your mother's side of your tree, and the same goes for your father's side when one of your paternal 1st cousins test. Of course, whilst all of your matches must be on one side or the other, you won't share all of them with one of your 1st cousins (though you will probably share all of the closest matches).
But knowing which side of your tree a match is on is only a start - although we only have 2 parents we have 64 great-great-great-great grandparents (which is about as far as autosomal DNA tests can reliably reach), so halving the number of possible lines from 64 to 32 doesnít get us very far. What you really need is 2nd, 3rd, and even 4th cousins to test - whilst the number of shared matches will reduce, the precision with which you can identify the shared ancestry will increase.
If you share a match with a 4th cousin the genetic cousin you've both been matched with must be descended from the 3G grandparents you share with your known cousin, or from one of their ancestors. And whilst you're clearly not going to pay for all your 4th cousins to test, the reality is that some of them will already have tested - which is why finding documented cousins (such as the 'lost cousins' you find through the LostCousins website) is so important.
When you want to target a particular 'brick wall', as you almost certainly will, it might be necessary to persuade someone who isn't researching their family tree to test - in which case you'll probably have to pick up the tab. Who should you pick? The ideal person would be someone who shares the 'brick wall' ancestor and no other ancestors - though in practice relatives like that are few and far between, because ancestors usually come in pairs. It's usually only when a half-cousin tests that you can identify which matches come from a single line.
If your 'brick wall' is a great-grandparent then you ideally need a 2nd cousin to test; for a great-great grandparent it would be a 3rd cousin, and so on. Sometimes you may have to compromise, of course - there may not be anyone in the right part of your tree who is both available to test and willing to do so.
Tip: some companies require you to give the name of the person who will be testing when you order the kit but Ancestry don't, which is handy if there's a sale on.
Earlier today it was revealed that the Registrar of Births, Deaths, and Marriages in Perth had arranged for the term 'Aboriginal' to be removed from records when they were digitised between 2007-15. According to an official "Other known words or expressions which may be regarded as offensive, disrespectful or hurtful include 'bastard', 'illegitimate' and 'incinerated' for cremated still-born babies".
A lot of things that happened in the past are offensive to us now, but Tippexing them out of history isn't necessarily the right answer. You can read more about this story here.
There are many British names - forenames, surnames, and place names - which donít follow the normal rules. Indeed some surnames are pronounced differently depending which branch of the family youíre referring to, and many place names with similar spellings are pronounced differently depending on the part of the country.
I don't know whether it is still in use, or if the dumbing down of our media has extended to the British Broadcasting Corporation, but the BBC Pronouncing Dictionary of British Names is an invaluable work, especially if you weren't born in Britain and exposed to the idiosyncrasies from birth. The pronunciation of names like St John, Cholmondeley, Featherstonehaugh, and Wymondham can be mystifying if one hasn't come across them before - when I started my first job after university in 1972 I answered the phone to a distinguished client whose surname was Chandos-Pole - well, I can't remember precisely what I wrote down, but it was nothing like that! For my surname it lists both the pronunciation I was taught to use by my parents, and the pronunciation that my father's Suffolk ancestors would have used.
My second-hand copy of the book was bought for 1p (plus shipping) at Amazon, yet it has a pedigree - it was previously owned by Tim Llewellyn, who was the BBC's Middle East correspondent from 1976-80 and 1987-92. I canít promise that if you buy a second-hand copy it will have a similar provenance, but buy it for its contents and you wonít be disappointed.
PS: let me know if, when reading this book, you discover that one of the surnames or place names in your tree is pronounced completely differently from how you expected!
Fake news is one thing but when a New York Times writer called a Yorkshire Pudding a dessert, all hell broke loose on Twitter as outraged Yorkshiremen and women rose as one to condemn the ignorance of the cookery writer. According to this BBC News article, which reported the furore, the first printed recipe for Yorkshire Pudding appeared in 1747, 29 years before the Declaration of Independence in 1776.
But honesty is the best policy - I have to confess that when I were a lad we finished off any left-over Yorkshire Puddings with Tate & Lyle's Golden Syrup. Mind you, we were lucky - unlike the reminiscing Yorkshiremen in the Monty Python sketch.
I don't know about you, but I find Pyrex casserole dishes incredibly useful - whether I'm cooking in the oven or the microwave, or storing left-overs in the fridge or freezer. But over time the handles of the dishes and lids get chipped, and whilst I haven't cut myself yet I decided it was time to treat myself to a new set.
I donít mind paying a fair price for something that's going to last, but nor do I object to snapping up a bargain - and to find that I could get a set of 3 Pyrex casserole dishes on Amazon for just £10 (supposedly reduced from £31.99) was a very pleasant surprise. So pleasant, in fact that I bought two sets! Best of all is that they've changed the shape of the handles, so they're less likely to chip.
Picking up bargains at the supermarket often leads me to cook dishes that I've never tried before. For example, I recently bought several packs of sliced spring greens and several packs of mushrooms at a fraction of the normal price - adding a packet of cooking bacon (full price, but still only 57p for 500gm) allowed me to serve up several delicious (and pretty healthy) cooked lunches which worked out at a mere 21p for each generous portion. Mind you, I've found that almost any food tastes good when stir-fried!
According to an article on the BBC News website a Japanese rail company has had to apologise because one of their trains left the station 25 seconds early. I can remember when it was a miracle if British train arrived on time - but one of most stressful experiences did involve a train that left early. It was around 20 years ago: my wife and I were on our way to Heathrow Airport to fly to the USA, laden with luggage, and when the taxi dropped us off at Stansted Mountfitchet station we were just in time to catch the fast train to London. Or so we thoughtÖ. in fact we arrived on the platform to discover that the train was already pulling out of the station a minute early - and since the next train arrived in London 45 minutes later we were in severe danger of missing our flight.
Fortunately the station was manned at that time, and when we explained to the gentleman in the ticket office what had happened he made a phone call. A few minutes later the Stansted Express train from the airport made an unscheduled stop to pick us up - what a wonderful way of putting matters right!
We've all got examples of where bureaucracy went wrong, but do you have any examples where they exceeded your expectations?
This is where any major updates and corrections will be highlighted - if you think you've spotted an error first reload the newsletter (press Ctrl-F5) then check again before writing to me, in case someone else has beaten you to it......
© Copyright 2018 Peter Calver
Please do NOT copy or republish any part of this newsletter without permission - which is only granted in the most exceptional circumstances. However, you MAY link to this newsletter or any article in it without asking for permission - though why not invite other family historians to join LostCousins instead, since standard membership (which includes the newsletter), is FREE?