Ellen Johnson and Mary Lyster – a DNA discovery

by Davina Best

For more than 25 years, Peter Sherlock has argued that Ellen Johnson was the biological sister of Mary Lyster. They were most likely the children of two convicts: John Lyster (d 1836), convicted in London and transported on the "Ganges" in 1796; and Margaret Mooney (d 1815), convicted in Derry and transported on the "Anne" in 1801. His research was first published in the HAFS Journal of May 2000. That article has now been republished as a post on this website.

With almost no new information about Ellen or her parents, and following our success in using DNA to identify the birth and parentage of Tom Winder, Peter and I spent several years scouring DNA matches for people who might point to Ellen's ancestry. Despite doing an mtDNA test of Helen Compagnoni - a direct female line descendant of Ellen - we still had no firm clues.

We then tried a different approach. Peter created a descendancy chart for Mary Lyster and her husband James Dodds, in the hope that we would find some living descendants who had done a DNA test or would be willing to do one. This showed that only two of James and Mary's children had living descendants: their eldest son James Dodds Jr, and their second daughter Ellen Dodds who emigrated to New Zealand where she married Henry Mahon Jervis.

Any living descendants of the Dodds or Jervis families would likely be about 5th or 6th cousins of the Winder and Hungerford descendants. While this is a stretch for autosomal DNA testing, there was precedent for this: for example, several Australian Hungerford descendants have DNA matches to descendants of Dr Richard Loane of Bandon, Cork, the grandfather of Catherine (Loane) Hungerford, who are their 6th cousins.

The task was especially difficult because there were very few people who had a family tree that had Mary Lyster or James Dodds (in Ancestry or other platforms). Perhaps the spelling variations and lack of records for John Lyster or Margaret Mooney had them stumped. If only they'd found Peter's article from 2000!

As Peter built out the descendancy lines, we searched DNA kits for people with relevant surnames. We sent messages to people who might know a cousin that had done a DNA test. But nothing worked.

Then, in May 2026, Peter found a person whose DNA matched to a couple of Tom and Ellen Winder's descendants. This person's online family tree included an ancestor called Ngaire Jervois, piquing our interest as Peter recalled that Mary Lyster had a great-granddaughter called Ngaire Jervis. Once again, we see the importance of the careful tree-building that goes hand in hand with DNA analysis! This find was really a bit of luck, the result of lots of work and due to the help of others.

At this point, we have to say thank you to those family members who volunteered their Ancestry DNA kit to us after last year's AGM. Peter came across this potential match with a descendant of Mary Lyster as he combed through some of the distant DNA matches of these volunteers, knowing their results might contain something new.

Despite our excitement, Peter's first thought was that this might be a random match, a crossover like we've seen before from the intermarriages of early colonial families. To be sure, the next step was to find matches from other lines descended from Mary Lyster. Ideally, we wanted to find matches from different lines of descendants of both Ellen Johnson and Mary Lyster.

We were fortunate that one of Ngaire Jervis's descendants agreed to give us access to their ancestry DNA test. Working through this test, and over a dozen other tests of Winder descendants, we hit the jackpot. We found DNA matches between descendants of three of Ellen Johnson's children: Jessie Winder (Simpson), Ellen Winder (Hungerford), and Thomas Winder, and descendants of two of Mary Lyster's children: James Dodds Jr, and Ellen Dodds (Jervis).

This was no coincidence.

The next step was for us to create a model in BanyanDNA - an online software application which works out the statistical fit when we enter all the known genetic distances gathered from Ancestry or MyHeritage. It identifies any anomalies that might be from intermarriages that we missed and allowed us to test out some unclear relationships.

The results to our two key questions from a quarter of a century ago were clear. Were Ellen and Mary biological sisters? Did they share one parent or two? We can now say with a high level of confidence that they were biological sisters and most probably shared the same two parents.

What use is this?

First, this information confirms that Ellen Johnson and Mary Lister were sisters, giving important context for Ellen Johnson's birth, parentage, and early life in the colony of New South Wales. This will inform Peter's forthcoming book, a biography of Tom White Melville Winder.

Knowing this gets us one step closer to finding clues to the identity and ancestry of Ellen and Mary's likely parents, Margaret Mooney and John Lyster. There is the possibility of identifying DNA matches to members of their birth families in Britain or Ireland, so the search continues.

We already have several "unlinked clusters" of remote DNA matches, groups that share a common ancestor who must be related to either Ellen and Mary's mother or father because the matches appear in tests of both Ellen and Mary's descendants. The most promising of these clusters is the descendants of William Harvey, an artillery driver in the British army who with his wife Margaret Graham and their many children emigrated to Ontario in 1822. The "Harvey / Graham" cluster seems a strong candidate for matching to Margaret Mooney, who was tried in Derry in 1799, because William Harvey and Margaret Graham were also from northern Ireland and were possibly married in Derry in about 1805. The search continues.

Can you help?

If you have taken an autosomal DNA test, please consider giving access to Davina Best and Peter Sherlock so we can check your shared matches to establish further connections to the families of Ellen Johnson Winder and Mary Lyster Dodds.

If your test is with Ancestry, update your online family trees to ensure that you are correctly showing the parents of Tom White Melville Winder as "William Winder" (1754-1803) and "Mary Norman" (1752-1827), and the parents of Ellen Johnson as "John Lyster" (d 1836) and "Margaret Mooney" (d 1815).

After a day or so, check the "Thrulines" tool to see if the software is showing DNA matches that descend from John Lyster and Margaret Mooney, and get in touch with us.

Through the combination of traditional archival research and DNA matching, we hope one day to establish conclusively the identity of our earliest Australian immigrant ancestors, John Lyster and Margaret Mooney, their places of birth and their birth families.

Descendancy Reports

For those who would like to compare the names of Ellen's and Mary's children, this first report below shows two generations descended from Margaret Mooney and John Lyster.

If you would like to look for Mary's descendants or possible Mooney or Lyster links in your DNA kits, contact us for a more detailed descendancy report.