Finding Roy’s Grandparent – How DNA testing helps your research


I have been involved in ‘genetic genealogy’ since I took a DNA test at 23andMe in 2012. Whilst I have found such testing interesting and occasionally useful, I’ve never used it for something “really important“. However, for the last few months, I’ve been working with a relative who has been searching for information about his ancestors. Information that would never have been found without the use of DNA testing. In this post, I want to describe the tools and techniques we have used to combine traditional and genetic-genealogy to identify an ancestor.

I should just mention that I have slightly changed the order of events that occurred to simplify the ‘story-telling’ of this investigation. In writing up this research I’ve used bold test emphasis either a name or a result.


Earlier this year I had a new and interesting DNA match on the website. This match was from a test taken by Roy Gress who lives in Australia. Roy was interesting for three reasons. Firstly, and most obviously his Gress surname was curiously close to my own, relatively rare, Grass surname. Secondly, he and I matched a known second-cousin, June. She and I are descendants of my great grandparents, George Grass and Elizabeth Ann Basham (more of them later). This DNA-connection implied that we were all related through either George or Elizabeth, or both, or one of their ancestors. Finally, my new match was showing, by the Ancestry website, as possibly being a 3rd cousin. Given that the Grass family is relatively well-researched, and that he didn’t seem to match any known Grass, my curiosity was well and truly piqued.

After some communication between the two of us, it became clear that our link must exist through his father, Frederick William Gress.

Traditional Genealogy

Roy has, for the past 20 years, been searching for information on his father’s parents firstly using the publicly-available UK records. However, it had reached the point that he was prepared to invest in a professional genealogical firm to research his family. This firm had found a number of records that covered much of Roy’s father’s early life and provided some suggestions on who his grandparents were.

Frederick William Gress

As with any family-history research, it’s important to understand what primary-source information is available i.e. the information recorded directly from the people involved, in this case, Frederick William’s parents. Roy knew his father was born on March 6th 1909. His birth record is shown below. It identified that he was born in St. Mary’s Hospital, Manchester. His parents were Frederick Gress and Isabella Gress, née Clarkson. They were living at 4 Clarendon St, Chorlton-upon-Medlock, Manchester. This was a slum part of a working-class district, more commonly known as ‘Little Ireland’ due to the number of Irish immigrant cotton-mill workers. Interestingly the birth is reported by his mother, 5 days later, on the 11th March. The father is rather wonderfully described as being a Theater Scene Shifter, which given the information on him we discovered later, rather aptly describes him.

Birth certificate for Frederick William Gress born 6th March 1909

Birth certificate of Frederick William Gress

The other primary-source record is his baptism record (see below). This has some significant differences to his birth record, but still enough detail to make it a believable match. Frederick is recorded as being baptized at All Saints Church in Chorlton-upon-Medlock on the 31st March 1909. This time his parents are recorded as William and May Gress living at 46 Clarendon St, which is close enough to match the earlier No. 4 Clarendon St. Normally I would put such an error down to poor recording by one of the registrars.

Baptism record from Frederick William Gress

Frederick William Gress, baptised 31st March 1909

What stands out from these records are two things. Firstly, the obvious, the fact that the given names of the parents are different. More subtly, the information about the father is rather vague. He did not report the birth, which is, in my experience, unusual. In addition, the given names that are reported for him match the two given names of his son, Frederick William. Was this just a coincidence or part of the invention? One thing is clear, there is no proof that the father was around at the time of birth (although the old adage that ‘absence of evidence is not the same as evidence of absence’ should probably apply). Naturally, at this point in the investigation, there existed the possibility that Frederick Gress was possibly an alias.

Normally in such a case, the obvious answer would be to look at other publicly-available sources to build a profile of the parents of Frederick William. Were they around in the 1901 or 1911 UK censuses and was there a marriage recorded for Frederick Gress and Isabella Clarkson? Using any combination of wildcard searches for Frederick Gress/Grass and Isabella Clarkson/Gress showed no marriage records. Either they married somewhere outside the UK, or they weren’t married (probably a not an uncommon situation around that time). Similarly, the 1901 and 1911 census returns provided no reasonable hints, and certainly nothing to show a Frederick and Isabella Gress living together with their son.

The professional genealogy firm that Roy had hired had done a good job of assembling both what the family knew and some harder-to-access documentation. Frederick William was actually raised in care. In the 1911 census, he is recorded as Fredrick McIntosh living with ‘foster parents’ in Waterloo, West Derby. Although the adoption was not recorded, it was not unusual for children to be “passed on” to relatives, friends or families more capable to care for their children. Later in his youth, spasmodic records (most were burnt in a record office fire) have him admitted to Fazakerley Cottage Homes, Liverpool, a children’s home housing up to 600 children in 24 cottages. The firm also provided some speculative information on the parents of Frederick William but these, in retrospect, don’t match the DNA evidence. He maintained the use of the McIntosh surname until he left the “child-care system” and started work, reverting back to his birth name.

The Grass Family

Before we go further it’s worth having a look at my Grass family roots. Given the closeness of the DNA match, and the 1909 date of Frederick William’s birth then the simplest solution is to look at the children of George Grass and Elizabeth Ann Basham. George Grass was born in 1864 in the village of Mundford in Norfolk. He was, as far as we know, the only son of John Grass and Charlotte Whittle and, by implication, the only recent Grass ancestor who could normally pass on his name to his children. Below are a couple of pictures of George Grass and his wife Elizabeth Ann Grass.

George and Elizabeth Ann Grass (nee Basham)


George Grass and Elizabeth Ann Basham were married in 1883 and had in total eleven children together, born between 1883 and 1906. This number is confirmed by the census records, especially the 1911 census, where married women were required to record the number of children she has had. Below is a simplified family tree. The last six children of the marriage are grouped as they are unlikely to have been the parent of a child born in early 1909. The oldest of these children died in 1908 and the next oldest would have been roughly 12½ years old at the time of pregnancy.

George Grass & Elizabeth Ann Basham family tree

a partial family tree for George Grass and his wife Elizabeth Ann Basham

If it helps, there is a rather nice photo of some of the family. It was probably taken around 1898, assuming the smallest child is Lilian Frances Grass (b.1896)

The Grass family circa 1898

The Grass family c. 1898
Back Row (L to R): Isabella May Grass, Lilian Frances Grass, Elizabeth Ann Grass (nee Basham), Agnes Maud Grass Centre Row: Margaret Elizabeth Grass, Arthur Headley Grass, Ethel Louise Grass Front Row: Daisy Laura Grass

DNA evidence

Ancestry DNA Results

It’s now important to look at how DNA testing radically changed our understanding of Frederick William’s parents. As I mentioned, Roy matched myself and a known second cousin – June. I matched Roy over 164 centiMorgans (cM) of DNA and June matched him with the slightly higher figure of 186 cM. To put this in context, the average match between two full siblings is around 2930 cM, first cousins around 870 cM and second cousins around 230 cM. The Ancestry website provides information on the possible connection between two individuals. Taking the slightly higher match between Roy and June, the following probabilities were calculated.

Ancestry,com relationship probabilities, based on a DNA match of 186 CentiMorgans

Ancestry,com relationship probabilities, based on a DNA match of 186 CentiMorgans

There is a similar service provided by the website, this site has the advantage of providing a visual representation of the possible relationships. As you can see from the statistics and chart the most likely relationship is somewhere around second cousins, with a small possibility of this being a third cousin relationship. relationship probabilities based on a DNA match of relationship probabilities based on a DNA match of 186 CentiMorgans – showing the possible relationships between two people sharing 186 CentiMorgans

Reviewing Roy’s matches

At this point in the research Roy, very kindly, let me have access to his DNA information on the Ancestry website. This is something I’ve never used before and something I wanted to treat carefully. Looking through his matches one thing became clear. Roy had other, more distant, DNA matches with people who descended from both his Grass ancestors and his Basham ancestors. Given what research has been done on both these families there is no known other relationship between these two families apart from the marriage between George Grass and Elizabeth Ann Basham. Taking all this information together implied only one thing. One of the parents of Frederick Williaam was the child of George and Elizabeth. The question would then be who?

What Are The Odds?

There is one other, extremely useful tool that can help in interpreting this sort of DNA data. This is the “What are the Odds” tool (WATO). As the website explains “WATO is designed to help you figure out where someone, called the target, might fit into a known family tree”. In this research, Roy is the target and there are three possible hypotheses. He could be the child of my father (George William Grass), the child of my cousin June’s mother (Margaret E. Grass) or the child of another Grass sibling. In the first two cases Frederick William Grass would have been a half-sibling to our respective parent. Putting the known centiMorgan matching values into WATO generates the following results. The scores in this chart indicate the relative probability of each event. The main conclusion is that hypothesis 3 (Roy is the grandchild of another Grass sibling) is “about 14 times more likely than the next hypothesis” (the next hypothesis being Roy is the grandchild of Margaret E. Grass).

screenshot from the “What Are The Odds” tool at

As an aside, if you are interested in the WATO tool, then I can recommend this lecture by Andrew Millard, one of the authors of the tool.

YDNA testing

There was one other test that we could use to narrow-down the parent of Frederick William – yDNA testing. Y-Chromosome DNA is the sex-chromosome DNA that is passed from father to son, virtually unchanged. If Roy was the son of a male Grass he should have the same yDNA as me. I’ve done a bunch of yDNA tests, including the FamilyTreeDNA BigY test. This meant there was quite a bit known about my yDNA. (You can see my yDNA if you look at the I-Y89241 variant in FamilyTreeDNA public Haplotree – see screengrab below). There are only two testers that are currently in my branch. Given this information, then the cheapest test to confirm/deny the hypothesis that Roy and I had the same yDNA was to test one of my current terminal SNPs. We chose to test the current ‘branch-defining’ I-Y89241 SNP via the YSeq testing organisation, which cost around US$20. The result showed that Roy and I had different results. This implied that the overwhelming probability was that he and I were not related on the paternal line.

I-Y89241 yDNA Haplotree branch

I-Y89241 yDNA Haplotree branch


After all the DNA testing was completed the ‘most likely’ scenario is that Roy is the grandson of either Isabella May Grass or her sister Ethel Maud Grass. At this point, genealogy facts end and interpretations begin. The two primary-source records for Frederick William’s birth and baptism state that his mother was either Isabella or May. It is not hard for me to believe that these names were “breadcrumbs” left by the mother, Isabella May Grass, to her son.

How the mother ended up in the slum district of Chorlton-upon-Medlock, Manchester is unclear. St Mary’s Hospital, Manchester was a progressive hospital specialising in nursing for women and children, so was perhaps a suitable place to give birth. Almost all of the Grass siblings started their working career ‘in service’ at one of the large private country houses that dotted England. The boys worked outdoors, sometimes following their father’s profession as game-keepers, whilst the girls worked as “domestics”. A not-unrealistic scenario was that Isabella was working at one such country house and fell pregnant, and was then sent to Manchester for the birth. What I do find surprising that she did not return to Yorkshire where her parents were living at the time. Isabella’s grandmother, Eliza Bowers, was born “out-of-wedlock” so it seems surprising that such an event would have had a stigma attached to it.

Can DNA testing provide more information?

One test I haven’t mentioned yet is mitochondrial DNA (mtDNA) testing. Mitochondrial DNA is the “maternal line” DNA that is passed from mother to child. Unfortunately, Frederick William died in 1995, so it is not possible to test his mitochondrial DNA, and his son inherits this DNA from his mother.

It would be possible theoretically possible to test another grand-child of Isabella May. This should show a slightly higher amount of shared DNA. This person would be a half first-cousin of Roy and would, on average, share  437cM of DNA compared with an average of 233 cM for a second cousin. The problem here is that these are average figures. The range of shared DNA for a half first-cousin is between 856cM and 137cM of shared DNA. A second cousin range is between 515 cM and 46 cM. This overlap would, in most cases mean that it would be impossible to say that the DNA proved anything.

This brings us back to finding traditional genealogical evidence, such as a letter or perhaps some records of Isabella’s employment that may finally close this case.

Final Thoughts

This is the first time I’ve been involved in helping someone else use their DNA test results. Personally I found it an incredibly interesting experience, however, I realise for my cousin Roy this is a much deeper issue. His father was adopted at birth, which led inevitably to issues later in his father’s life. This lack-of-knowledge of his heritage was also, naturally a personal issue for Roy. He wanted to give, not only himself but his Grandchildren, an ancestral history, so they would know their family origins.

For those people, like myself, who grew up with more-complete knowledge of my ancestors, at least to my grand-parents level, it is a challenge to realise the effects this has on a person’s life.

Being involved in DNA research on someone other than myself also raised a number of issues regarding the ethics of working with someone else’s DNA. I plan to write a separate post covering this.

Having “completed” this research work, as much as Roy and I can, it also raises a much more general question – when do genealogical evidence end and belief begin. I feel confident that Isabella May is Roy’s grandmother, based on the evidence available. You could argue that this is only a belief, rather than proof, however, there are many things in genealogy that are believed. If you consider that average non-paternity rates around the world are believed to be between 1% and 2% then even the most solid genealogical paper trail will eventually have errors. On top of this, there are inevitable issues with the interpretation of what genealogical records do exist. In the end, every family-tree consists of some amount of belief.

Finally, there is the question of what further steps can be taken to identify Roy’s paternal grandfather. The first obvious step is to keep “fishing” – in this case, fishing for a close DNA relative who doesn’t match relatives of any of Roy’s other three grandparents. There may be a family story out there that matches the facts of Frederick William’s birth. As part of this genealogical “fishing trip”, it would be possible to transfer Roy’s DNA results to some of the other major genetic genealogy companies, such as MyHeritage and FamilyTreeDNA. This he intends to do in the not too distant future.

The other step that Roy can take is more complete yDNA testing, typically the FamilyTreeDNA y37STR test, which he intends on doing. There is no guarantee that this will identify his grandfather, but it might at least suggest a surname to further research

This entry was posted in Genealogy, Genetic Genealogy and tagged , , . Bookmark the permalink.

4 Responses to Finding Roy’s Grandparent – How DNA testing helps your research

  1. patc44 says:

    Thank you for sharing this information, how interesting our family heritage is! I use articles like this to help me unpick my family through DNA and understand the available tools out there. Being adopted throws up a totally different level of complexity! I find the new Beta version of Ancestry very helpful – to a degree. My understanding of the other tools available is limited, and difficult to grasp. But I will plod on.

    • Phil Grass says:

      Good luck with your search, I can only imagine how hard it is (emotionally and logically) doing such research. Whilst it takes some time to get your head around it, I suspect the WATO tool will be useful to you. In this case you will be the unknown person and need to fit yourself in various places in the family-tree to see which is the most likely relationship.

  2. Pingback: The DNA-Secret that went to the grave too early | learnalittleeveryday

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.