DISCUSSION OF ENGLISH DNA MARKERS

Return to Opening DNA Page        Updated Sept. 2012           Return to DNA Testing Results Page
Understanding Matches
The U.S.A. Markers Discussion

(Additional discussions of some of the following topics are found in issue 55 and later issues of the Journal. The U.S.A. participants’ DNA results are discussed in a separate article accessed by a link above. For this discussion your editor relied heavily on data previously supplied by Timothy Mangham, especially for the lineage information)

DNA AND THE ENGLISH SURNAMES

We picked 400 years ago (about 1600 AD) because that is about 100 years (in round figures) before John Mangum came to the New World². It would have been extraordinary luck if we ever find a DNA participant who is a descendent of the parents that John Mangum was born into. More than likely, if we do find a participant with a close relationship to the U.S. Mangums, his lineage may link up to the immediate family of John Mangum through a common ancestor 100 or 200 years before the time of John Mangum [who may have been born by 1674³]. Even this would still require a bit of luck on our part. Take my own “Mangum” lineage. Few of the Mangums reading this article have a common ancestor with my Mangum lineage after the time of Howell Mangum who was born about 1750 (250 years ago) and some do not join my lineage before we get back to John Mangum the immigrant who was born about 1674 (almost 330 years ago).

Of course the lineage spread in 17th century England might not have been as extensive as it is now in the U.S., but allowing just 100 years before 1692 for our “parent family” participant (if we ever find him) is asking for a lot of luck. We might want to consider that we have found the parent family of John Mangum if we find a significant chance of a MRCA at about 600 years or 1400 AD. That is only 292 years before the time that John came to the Virginia Colony. Maybe this could be considered a partial success if we can convince ourselves that the relationship shown by DNA testing is real and not a chance match within a common Haplogroup⁴. We have little chance of finding relationship before 1400 AD because surnames were just coming into common use. Without surnames, searches in what few records exist from that time would essentially be useless. Also, our genetic Mangum kinfolks who lived much before 1400 AD would have left descendents with a variety of surnames and only a small percentage would have become Mangum or whatever English variant of the surname our immigrant ancestor’s family used.⁵ Although we will almost certainly never be able to find these very distant relatives of pre-1500 AD by traditional genealogical means, DNA analysis may give us tantalizing hints of their relationship to us.⁶

DNA test results are statistical, meaning that they do not show absolute relationships between individuals. All results are expressed as a percentage likelihood of a MRCA at a certain number of generations in the past. In other words, it could be expressed as “10% chance that person X & person Y shared a common male ancestor no greater than 10 generations in the past.” On the other hand that same example also infers that there is a 90% chance that the common ancestor is greater than 10 generations in the past. Since generations are assumed to be 25 years on the average, the time to the MRCA can be expressed in years. 10 generations would be 250 years in this case. Typically when two people match 12 for 12 markers there is a 50% chance that they shared a common ancestor no greater than 7 generations ago (175 years ago). With a 25 for 25 marker match the time to the MRCA is only 3 generations, again at the 50% confidence level. Matching all 37 markers reduces the time to the MRCA to 3 generations or less. We are now testing for 111 markers. A 111 marker match will reduce the time to the MRCA even further, probably to immediate family groups. Unless the comparison is between immediate family groups (father, son, brothers, uncles), there will most certainly be some mutations, that is, not an exact 111 marker match. For the four of us who have upgraded to 111 markers, there are a few mismatches (mutations).

A 50% probability of a common ancestor may not seem to be very confident and it does leave open a 50% chance that the common ancestor lived more than the calculated number of generations at the 95% confidence level, meaning an almost certainty that the time to the MRCA is correct. At the 95% confidence level, there is only a 1 in 20 chance that it is not as calculated. As you might expect, the number of generations to the MRCA increases significantly as we increase the confidence level but it decreases as the number of marker matches increases. An exact 12 marker match with a 95% confidence level places the time to the MRCA at up to 29 generations or about 725 years. At 25 markers it is 13 generations and at 37 markers it is 7 generations.⁷

As a practical matter we are more often dealing with a certain number of mismatches between individuals. A mismatch will always increase the time to the MRCA, and as the number of mismatches increase, the likelihood of a relationship within the time in which surnames have existed decreases accordingly. That span of years in the past in which surname existed is of course about 600, or about 1400 AD.⁸ The calculation for the increase in ‘time to the MRCA’ per mismatch is highly dependent on the rates of mutations for each specific marker involved.⁹ Laboratories in the past have been rather secretive about what values they place on these mutation rates, as determined by their research, but they did offer averages. They are just now beginning to offer estimated mutation rates for some of the markers they test for, and these markers rates are included in the above tables. Fortunately, FamilyTreeDNA calculates the times to the MRCA for you, based on specific marker mutation rates, and give you that value for all combinations of persons in our surname group.

With the above background under our belts, it is time to discuss the relationship between the U.S. Mangums and the English surnames.

MANGHAMS
The Manghams are fairly common in England, at least as compared to most of the other related surnames we have studied.¹⁰ They are concentrated in Yorkshire & Lancashire. In fact, most modern day Manghams are situated in West & South Yorkshire in north central England, although there are scattered Mangham families throughout England.¹¹ It may be that the English Manghams are not one single genetic group, but may have descended from several different ancient surnames. However, three of our four English Manghams have identical DNA. One theory, which now seems to be discounted by our research, is that the Manghams descended from the Manninghams, from the ancient village of Manningham. Other theories have the Manghams descending from Mallum (Malham), Malgham, Mangall, etc. Our DNA results so far has failed to find any significant relationship between the English Manghams and any other related spelling surnames.

The 1881 census of England shows 219 Manghams and 74 Manghans. Most are in the north of England, Lancashsire and West Riding¹² of Yorkshire plus a few in Durham and the rest scattered around England.

Our DNA tests have shown that the English Manghams are totally unrelated to the U.S. Mangums, at least for the families of the 4 individuals tested. The English Manghams are in a totally different Haplogroup that came from the Mediterranean area. Any relationship they have with our Mangums or any of the other surnames in our Mangum DNA group, including the English/Irish surnames, must go back thousands of years when Europe saw extensive migrations from the Mediterranean, Middle East and Asian areas. Another possibility is that the Manghams descend from Roman soldiers who occupied England from 43 AD to 410 AD. This would explain the Mediterranean origin, however, there did not seem to be extensive intermarriage between the Roman invaders and the local population although it surely happened on occasion.

Of course there is still a chance that other Manghams from different extractions might be the parent family of the U.S. Mangums. At this point however, we are no longer recruiting English Manghams for testing.

MAUGHAMS_MAUGHANS
The Maugham is another fairly common English surname and seems to be a toponymic, or a name which was taken from that of a specific town or village. In this case, it may have been adopted from a village in the Strathclyde region of southwestern Scotland, near the city of Dalserf in Lanarkshire. The original spelling was Machan (pronounced 'mawn'). Later the family moved into Northern England and into Yorkshire.

This spelling of the surname is found mostly in the areas around Newcastle and Bishop Auckland in the north-east of England. The descendents of the famous writer Somerset Maugham believe their Maughams may have originated in the West Riding of Yorkshire.¹³ They also opened the possibility that the Maughams were related to the Malhams of Kirkby Malham although our DNA tests do not support that belief.

There were 436 Maughams in the English 1881 census. 149 were in Durham, 44 in Lancashire, 27 in Cumberland (Cumbria), 43 in West Riding of Yorkshire & 42 in Northumberland, all in the north of England. London (Middlesex) had 34, Scotland had 1 and Wales had 3. The rest were scattered about England.

DNA testing shows that three of the Maughams have very similar DNA, and obviously had a fairly recent ancestor. Traditional genealogy shows all to have originally had the Maughan surname. The Maugham without a known Maughan ancestor (Kit 16470) does not seem to be related to the Maughans.

In the 1881 census there were 1672 Maughans, most in the north of England, Durham, Northumberland, and the Ridings. There were a few in Scotland and Wales. DNA shows that the Maughams and Maughans are not related to the U.S.Mangums.

MANGLES
We presently know very little about the origins of the Mangles. The ‘Mangle’ surname in the 1881 census was concentrated in the northern part of England in Lancashire and Yorkshire (West Riding) [38 individuals]. One, two and three individuals were found each in Berkshire, County Durham, Jersey & Sussex. 20 were found in Scotland. There were 66 individuals total. There were a total of 123 ‘Mangles’ in England in 1881. Most were in Yorkshire (West Riding) [46] and County Durham [32]. 15 were in North & East Riding of Yorkshire, 5 in Sussex, 6 each in Lancashire & Hertfordshire, 7 in Middlesex (Now London) and 1 to 3 in other Shires. None were in Scotland. No census information for Mangle or Mangles in Ireland was available.

Our Mangle participant had a 5 marker mismatch with the Mangums at the 12 marker level and 10 mismatches at the 25 marker level. FTDNA website shows no significant matches between this participant and anyone in our Mangum surname group.

MANNINGHAMS
The Manninghams seem to have originated from the village of Manningham, which has now been incorporated in the city of Bradford in West Yorkshire. The village was probably an old Anglican settlement from the 7th century. The Manningham surname may have evolved from the Old English personal name Maegen. English surname experts believed that the Mangham surname derived from Manningham. However, our DNA testing results seems to cast some doubt on that theory. The 1881 census of England has only one Manningham listed, living in the West Riding of Yorkshire. [See the Manningham article in the Journal issue #54, page 4.]. One of our Manningham participants speculated that the lack of Manninghams during this time period was because they were forced to immigrate to other English Colonies, possibly even to North America. Those individuals who assumed the village name of Manningham as their surname most likely came from many different genetic families.

Initial tests (12 markers) with our two Manninghams were very promising. There was only a 2 marker mismatch with the Mangums with one and a 5 marker mismatch with the other, but surprisingly both showed a 17-18% chance for a common ancestor with the Mangums at 400 years and a 35-37% chance at 600 years.

The Manningham’s results were close enough to the MML to warrant further testing. The 25 marker test for that one of the Manninghams was even more encouraging, increasing the percentages to 23% at 400 years and 53% at 600 years. Unfortunately the 37 marker test was very disappointing. There were 5 & 6 more marker mismatches in the 26-37 markers. The percentages for a recent common ancestor with the Mangums dropped to insignificance.

The real surprise is that the two Manninghams do not seem to be related to each other!

MENGHAMS
The Menghams and the Minghams that follow below all seem to have originated on and around Hayling Island in Hampshire, on the Southern coast of England. There is a small village called Mengham on the island. Both Mengham & Mingham were used interchangeably by the families in the area. The 1881 census shows that most of the Menghams [27 of 31] were in Hampshire where Hayling Island is located.

Our Mengham participant had 6 out of 12 mismatches with our Mangums. The lab calculations give a 1% chance of a common ancestor at 400 years and a 5% chance at 600 years. The Mengham does seem to have a significant genetic relationship with the Mingham below

MINGHAMS
As noted above the Minghams probably originated on Hayling Island and are found in several areas in Hampshire around the island. The 1881 census shows that almost half [10] of the 21 Minghams were in Lancashire with only one in Hampshire. The rest were scattered around England with 3 in Scotland.

Some of the Minghams became ship captains and were involved in shipping to Jamaica about the time John Mangum came to the New World. There are several other records of Minghams in Jamaica around this time. This may be just coincidence, but it interesting nevertheless. Note also that John Mangum was listed as a Mingham in a 1695 deed in Virginia. In July of 1724 a Benjamin Mingham was transported to Virginia from Hampshire, England as punishment for some minor infraction of the law. We have found no New World record of Benjamin Mingham.

See the Journal article on the Menghams & Minghams in issue 50, page 29.

The Mingham participant had 4 mismatches in 12 with our Mangums, meaning a 5% chance for a common ancestor at 400 years and a 14% chance at 600 years. The Minghams have a 10 for 12 marker match with one of our new Irish Mangan participant, but it is difficult to know if this is significant. See more with the Mangan discussion below.

MALHAMS
There were 95 Malham individuals in the 1881 English census. Most resided in three counties; Berkshire (17), Surrey (22) & the West Riding of Yorkshire (27). County Durham had 9 and Middlesex had 6. Midlothian in Scotland had 1. Berkshire & Surrey are in the South Central of England, County Durham and Yorkshire are in the North. Middlesex is no longer a shire but was in the area now taken by London. [See the Malham article in Journal #45, page 21.] Our Malham participant does not seem to be related in any significant way to anyone else in the database. There is a 16% chance that he is related to the Maughams, the Minghams and to our Ellison G. Mangum descendent within the past 600 years, a quite low percentage. [Note: We already knew that this Ellison G. Mangum descendent is not related to the main line of Mangums.] The family of the novelist Somerset Maugham considered the Malhams as a probable procurer of the Maugham surname. Our DNA research shows that scenario unlikely, but barely possible. The Malham’s relationship to the main line of U.S. Mangums is less than 5% at 600 years.

MINGS
We have a descendent of Johan Jacob Ming who was born 1727 in Rotenbach, Wurtemberg, Germany. Johan’s descendent did an independent DNA test, not using any surname group. Because of the spelling of his name the lab suggested he join our surname group. Your editor concurred and he became a member of our group. Unfortunately he does not match any of the others in our surname group. He is in the “E3b” Haplogroup. This Haplogroup is believed to have evolved in the Middle East. It expanded into the Mediterranean during the Pleistocene Neolithic expansion. It is currently distributed around the Mediterranean, southern Europe, and in north and east Africa. This individual has since left the Mangum DNA surname group.

The closest match to any of our Mangum group participants is with the English Manghams. He is 4 markers in 12 away from several of the Manghams. Lab calculations show an 18.83 chance of a common ancestor at 600 years in the past. We conclude that the chance of a real genealogical relationship with the English Manghams is very small.

MANGAN
Timothy Mangan, great grandfather of one of our Mangan participants, was born in Bally Bunion County Cork, Ireland about 1880. He and his brother moved to England in the early 1900's. He met his wife there (who has not been identified) and their first child was born on 23rd February 1904. This child died 27 Dec. 1971. We know that Timothy died in the war (WWI) as his name is in the town hall with all other soldiers that died in the war. We don't have any information concerning the family of the other Mangan participant.

The Mangans did not match any of the U.S. surnames, but one had a distant match with the English Mingham. The match was 10 out of 12 markers. FTDNA reports that there is a 40% chance of a common ancestor between the Mangans and Minghams in 24 generations. I suspect that this is not an indication of a true relationship. The Mangan did have numerous exact 12 marker matches with other surnames (about 50), but such matches at the 12 marker level are usually not significant. Probably the 10 out of 12 match between the Mangans and the Minghams is simply an accidental convergence.

In spite of the “Random Sampling” analysis, we still should not totally dismiss the English surname results. The 12 marker test was never meant to actually determine whether two participants have a close relationship, even when the surnames are identical. It certainly cannot determine close relationships when the surnames are different. As an example, this writer’s first twelve markers (which are identical to the Standard Mangum Markers) shows perfect matches between this writer’s 12 marker DNA markers and those of 11 participants in other surname groups which had differently spelled surnames. Many of these other participants were contacted and none of them could find any possible relationship to the Mangums.

Because even perfect 12 marker matches within a particular surname leaves a lot of doubt about how close the relationship is, we usually go to 25, 37 or even more markers. The value of the 12 marker test is that it can often show that two individuals do not have a close relationship. When two individuals are found to be in different Haplogroups¹⁴, or if there are more than 5 or 6 marker differences, we can almost always discount them as having any relationship (common ancestor) within a genealogical time frame, that is, within the last 600 to 700 years. When the number of mismatches is less than 5, and the surnames are identical, or of similar spelling, we begin to allow the possibility of a relationship, but it is significant only at several hundred years in the past. Take a look at the following graph. It shows how the confidence level (in percent) for a relationship (common ancestor) decreases as the number of mismatches (out of 12 markers total) increases. The graph charts the confidence percentages (for a common ancestor between participating individuals) at 600, 400, & 200 years in the past when there are 3, 4, 5 & 6 mismatches. Note that at 5 mismatches the confidence level for a relationship even at 600 years in the past is very low. At 3 mismatches there is a low but significant confidence of a common ancestor (15%) at 600 years.

Important note: These confidence levels for a common ancestor apply only for identical or almost identical surnames (like Mangum, Mangrum & Mangham), and to a lesser extent for surnames with only general similarities. For totally different surnames, especially with the common R1b Haplogroup, these confidence levels do not apply. For different surnames, even exact matches at the 12 marker level do not normally indicate a significant genealogical relationship¹⁵. The values in the graph above have been extracted from Family Tree DNA Lab data, and should be considered as only approximate values.

There are several other English surnames (including Mangall) we would really like to see results for, and several Irish surnames. Hopefully some of these will eventually be tested, but seeking out participants is presently not practical.

NOTE: Since these tables were generated, we have had several new English/Irish participants. They do not really change the information in the above tables. We have four Mangham participants instead of two. Three of those have identical DNA at the 12 marker level and ManghamI represents these three. Likewise, we have four Maughams instead of two, but MaughamI represents three with identical DNA at the 12 marker level. The only real difference is the additional Irish Mangan, but he has no real matches with the Mangums or any of the other Irish/English surnames.

Some of the surname pairs show low but significant levels of confidence that the surnames had common ancestors at 400 & 600 years ago. However, some surnames seem to have a much higher or lower percentage. The Manghams had extremely low levels of confidence of a common ancestor with any of the other surnames, including Mangum. The two Manghams had zero confidence in a common ancestor between them at 600 years. The Mangles have very low confidence for a common ancestor with the Menghams & Minghams. Not surprisingly the Minghams and Menghams have a moderately high confidence of a common ancestor between themselves. Present theory is that both originated on Hayling Island and are generally thought to be related through a relatively recent common ancestor. The Maugham has a moderately high confidence of a common ancestor with the Mingham, but only a low confidence with the Mengham, a somewhat paradoxical result since the Minghams and Minghams seem to be related.

The match between one of the Manninghams and one of the Maughams was the highest in any other pairs. At the 600 year level, the confidence of a common ancestor was almost 70%.However, at the 37 marker level, the confidence dropped to 0.6% at 400 years and 7.5% at 600 years. This just shows the danger of relying only on the 12 marker test. The chance of an accidental match between unrelated individuals at the 12 marker level is very high.

The following pairs of surnames were found to have low confidence for a relationship at 600 years ago: Mangham/"Any other surname", Maugham/Mengham, Mangle/Mengham, Mangle/Mingham, Malham/Manningham, Malham/Mangle, Mangan/Malham and Manningham/Mengham. We would expect that testing for more markers would clarify these relationships somewhat. Testing at 12 markers is considered marginal in determining genealogical relationships and testing for more markers will often completely invalidate the 12 marker results.

As with most of the surname comparisons, the Mangums have a low but significant level of confidence for a common ancestor with most of the English surnames at 400 & 600 years. The comparison with the Manningham is significantly higher, but unfortunately it drops back into the low but significant range when comparing all 37 markers.

You may be wondering why a certain number of mismatches between various pairs of surnames does not always give the same percentage of confidence for a common ancestor. That is because each marker has its own rate of mutation, which is calculated by the lab. This translates into different confidence levels when the mismatches are at different markers.

In order to try to answer those questions your editor looked at five different R1b Haplotype surnames at random on the Internet, all totally unrelated to Mangums, and examined 12 marker DNA results from several individuals in each surname. There were 20 individuals in all. The average DNA marker mismatch between the surnames and the Mangums was 3.5 markers with a range between 2 & 5.

For the Mangums, this DNA survey shows that the related spelling English surnames are really no closer to us than random R1b Haplotype surnames. The Manningham could be considered closer than random, but as discussed above, the higher resolution tests showed that the 12 marker tests were a fluke. In considering the entire 37 marker panel we can only conclude that the Manninghams are also no closer to our Mangums than random surnames.

The matches between the Menghams/Minghams seem to be closer than random, as does the Maugham/Minghams, the Mangan/Minghams and one of the Mangan/Manninghams. The other Mangan/Manningham has a low probability. The rest are no closer than random surnames to each other. Again, using 12 markers means low resolution and higher number of markers could completely change this evaluation.

Even exact twelve marker matches between different surnames can almost always be dismissed as accidental matches. Only when the surnames are the same, or have generally similar spellings can we assign significance to the matches. Whether we can assign any significance at all to the above, which have several mismatches at the 12 marker level, remains to be seen.

Marker #439 is another interesting marker. Most of the U.S. surnames have 14 here. Exceptions is the Isom Green Mangrum participant and one of the descendents of Joseph Mangum (#50823). Those two have 13 at this marker. All the English surnames (except the Manningham who also has 14) have 12 or 13. We presently do not know if these marker similarities or differences have any significance.

It is still possible that we will eventually stumble on an English, Scottish or Irish DNA participant who has a surname spelled similarly to Mangum, and who has a close match with the U.S. Mangums. However, I am beginning to believe that scenario unlikely. It is more likely that we will find a high resolution close match with an entirely different surname, if we find anything at all. We will probably never be able to determine the actual relationship of any of those differently spelled surnames when all we have is a low resolution match. John Mangum’s DNA was Haplogroup R1b1, a very common Haplogroup in England. We have already found dozens of low resolution DNA matches with him, and we have no way to determine if the relationship is real or much more likely a relationship that goes back more than 600 years, or maybe even an accidental match through convergence.¹⁸

______________________________________
FOOTNOTES
1. MRCA stands for “Most Recent Common Ancestor”.
2. Actually only about 92 years as John Mangum was in the Virginia Colony by 1692.
3. Assuming John was 18 years old when be first left tax records in the Virginia Colony.
4. Our Haplogroup R1b is a very common and chance matches across differently spelled surnames, especially at the 12 marker level, are also very common. However, because our matches are between similarly spelled surnames, the matches are deemed to be more significant than matches between non-similar surnames.
5. The time frame for the adoption of surnames depends on the location of your ancestors in the past. For example, in England, the major period of the formation of English heredity surnames began in about 1250. Surnames have changed considerably in form over the centuries, and many variants arose, of which not all survive to the current day. The adoption of hereditary surnames was a slow and irregular process. Some rich Londoners possessed hereditary surnames by the second half of the 12th century. In the countryside, the idea of hereditary surnames took longer to take hold. By about 1350, over half the rural families had firm surnames.
6. Few genealogical records exist in England before the 1500’s
7. Personally I tend to believe that these calculations are very conservative and the probabilities for a common ancestor are much higher than shown above. Offhand comments by the scientists involved in DNA testing projects seem to confirm this.
8. There is no fixed date for the beginning of surnames. It was an ongoing process for hundreds of years. Royalty, the aristocracy and estate holders used them first. The peasants and common man had little use for them until later. Surnames also developed at different times in different areas of the world.
9. Rates of mutations of markers vary considerably, some maybe two or three times the rate of others.
10. The Maughams are the most numerous, at least according to the 1881 census of England.
11. Although the Mangum name is almost unknown if England today, it did once exist in the Yorkshire region.
12. Before 1974 Yorkshire was divided into three “Ridings”, West, North & East. In 1974 the Ridings were discontinued.
13. The book they wrote, Somerset & all the Maughams by Robin Maugham, was published in 1967.
14. The Haplogroup is determined by a subset of the first 12 DNA markers.
15. Of course, we are all related at some point in the past.
16. A lineage break is defined as an adoption, name change, extra-marital event or any situation in which a male child does not carry the genes associated with the surname he uses.
17. The Irish surname research will be much more difficult than the English research. In the 1920s the Irish government decreed that surname spelling would be consolidated. In the process many similar spelled surnames were combined into a single spelling. Many combined surnames actually had no relationship with each other.
18. Surnames for the common people began to come into existence about 600 years ago, and the spelling of adopted surnames at that time was fairly random. Close family members would not necessarily adopt the same surname. Convergence is something else. DNA markers change slowly over the years, and occasionally these random changes in unrelated surnames will accidentally converge, creating a low resolution match. This occurs often in common Haplogroups like R1b. This is what we call convergence.