This is a Message Board Post that is gatewayed to this mailing list. Author: ksaxe Surnames: McConnell, McDonald, McDonnieal Classification: queries Message Board URL: http://boards.rootsweb.com/surnames.mcconnell/2482.1/mb.ashx Message Board Post: Finding and Understanding "Mismatched Matches" If John C. and Robert A. McConnell had not known that they were related prior to their DNA testing, it really would not have been very hard for them to find each other. The only tricky part would have been to understand that they might be quite closely related, even though there are 6 steps between their 37 marker results, and they are not classified as 37 marker matches for each other by FTDNA. This is a case where a little bit of knowledge might be a dangerous thing. Normally a 25 marker match that "falls apart" at 37 markers doesn't indicate a common ancestor during the last few hundred years. When new results are added to the FTDNA database, e-mail notifications of 12, 25, 37, and 67 marker matches are sent to matching parties so long as both parties have agreed to have their information shared with their matches. The matching parties can also find their 12, 25, 37, and 67 marker matches listed on their personal match pages. If a man matches at a certain level with another man who has had more markers tested, the number of markers tested for that man will be listed next to his name. So with e-mail notifications of their 25 marker match, no notifications of a 37 marker match, and the availability of 37 markers for comparison noted on the match page, John and Robert would wonder what their match meant and why it seems to "fall apart" at 37 markers. Since they were tested at FTDNA, it would make sense for them to visit FTDNA's page on understanding 37 marker matches, which states that 33/37 marker matches are probably related. The moral of the story is that you should check out your McConnell surname variant 25 marker matches, even if they aren't listed as 37 marker matches on your match page. You should contact these matches and compare your results on all available markers. You may find the use of the project tables and/or ysearch.org handy for the comparison of results. If you match on 31 or more of 37 markers, you may share a common ancestor since the adoption of surnames. The probability of such a relationship with a 31 marker match is so low that it is noted only in the fine print. 32 marker matches are classified as possibly related. Matches at the level of 33 markers or above may be classified as mismatches if there are multiple step differences at one or more markers. If you find one of these near matches, y! ou should contact Doug McDonald and/or FTDNA to help you with statistics that can help you assess the probability of a relationship. It's generally recommended that you try to find other persons with test results between yours and your near matches in cases like these. The existence of such matches would suggest a cluster due to a common ancestor whose Y chromosome was probably most similar to the Y chromosome or chromosomes of any men in the middle of the cluster. Men further from the center would probably have experienced recent mutations differentiating them from the rest. Most of these near matches would probably match you at both 12 and 25 markers, allowing you to identify them and begin the process of making comparisons. As always it's a good idea to check databases like ysearch.org and the SMGF database for matches. When using Ysearch, it's good to start out with a genetic distance of 6 on 37 markers. If you have results for 25 markers total, you should at least take a look at 22/25 matches using the same methods. 21/25 matches are extremely unlikely to share a common ancestor since the adoption of surnames, but you should take a look at any 21/25 matches bearing a McConnell surname variant. Another way to find your near matches within the project and easily see your mismatches is to use the comparison function on the project tables since differences from the comparison person are all highlighted on the display. The relationship trees posted on the project website seem to be especially helpful with some of these near matches. Sometimes they show dots for two people very close together on the tree, but on different branches, suggesting that results for their lines have converged. This can happen when both lines experience the same mutations or when mutations in the lines bring them closer together. When this has happened, the men with the convergent lines are not as closely related as a 1 to 1 comparison of their results would suggest. In other cases, the opposite will be true. Two people will be on the same branch of a tree, fairly close to a branch point, but their dots will be relatively far apart due to suspected recent mutations in at least one of the lines. These people will probably be more closely related than a 1 to 1 comparison would suggest. John was tested rather recently, and his results don't appear on the trees, but you can find a dot for Robert on both of the trees. It's interesting to me that the dots for most of the men who appear on both trees seem to have the same neighbors on both of them. This isn't true for Robert, which seems to show how the results for the 26-37 marker panel change the picture for him. There are some other interesting features of the mutations in their lines that I will share with you another time. In the meantime, it's interesting to note that L. J. McConnell on the R1b Yellow table has a similar match with S. H. McDonnieal. They match each other 33/37, with a 3 step mismatch on marker CDY B. These men are not listed on Ysearch, so I have not contacted them and know nothing about their family histories. They probably aren't as closely related as John and Robert, but their match is definitely worth exploring. Important Note: The author of this message may not be subscribed to this list. If you would like to reply to them, please click on the Message Board URL link above and respond on the board.