I just sent this message to the FTDNA HelpDesk ( [email protected] ): I'd swear it used to say on our member pages that you were going to *merge*, not *purge*, the separate Paternal and Maternal GEDCOMs attached to our accounts by the end of the year. I'm distressed to see this change because it means deceased members without living contacts will then lack GEDCOMs, entirely. This is a dreadful loss of valuable information. Test results without lineages are virtually useless. I urge you to *merge* the existing GEDCOMs, not purge any that are still existing on the cut off date. I fully realize how much more work that is for you, but what you are proposing is a bad move for your clientele. At the very least, if the subject is male, save the Paternal GEDCOM; likewise, if the subject is female, save the Maternal GEDCOM. Or, you could just leave well enough alone. Diana
Sounds like they've fixed what was wrong with the upload mechanism because that's what I wasn't able to do, before: that is, just change the "root" person in the file and have it work. By the way, if you still have two GEDCOMs uploaded, note that by the end of the year, these files will be purged from your account. You need to upload a combined GEDCOM by that time. I rather wish they wouldn't purge these old GEDCOMs because there are deceased individuals who will then have no GEDCOM, at all. The note on our member page used to say they would *merge* the files by the end of 2012, not purge them. I think we ought to pressure FTDNA to go back to merging them. I can see why they don't want to merge them - it's much more work for them. But it's their bad planning that they created this situation, so I think they should bite the bullet and take the harder, but better solution. Diana > From: Jim Bartlett > Sent: Tuesday, May 15, 2012 12:25 AM > > I loaded my GEDcom at FTDNA today and the software asked me which > name went with the DNA - so I highlighted my name (number 1 on the > list) and it correctly loaded my 12 generations. I the used the same > GEDcom for my father's account, and this time highlighted his name, > and the software selected the correct 12 generations that go with > him. Amazing! I had been dreading spending the whole evening > deleting half my GEDcom for his upload - what a treat for that to be > done for me by one click.
I loaded my GEDcom at FTDNA today and the software asked me which name went with the DNA - so I highlighted my name (number 1 on the list) and it correctly loaded my 12 generations. I the used the same GEDcom for my father's account, and this time highlighted his name, and the software selected the correct 12 generations that go with him. Amazing! I had been dreading spending the whole evening deleting half my GEDcom for his upload - what a treat for that to be done for me by one click. Jim - Sent from my iPhone - FaceTime! On May 14, 2012, at 12:08 PM, Walter J Freeman <[email protected]> wrote: > On 5/14/2012 4:40 AM, [email protected] wrote: >> I would need separate GEDCOMs for my parents who were from 2 different >> parts of the world. Their ancestral paths didn't cross. Therefore, it is >> necessary to concentrate heavily on the geography. >> >> >> --Ida Skarson McCormick, [email protected] >> >> In the begining, FTDNA required separate GEDCOMs for the maternal and the paternal line. This is no longer the case. In fact, it is much easier for all, if one just uploads an Ancestor tree GEDCOM so that both maternal and paternal lines are displayed. > Like you, Ida, my maternal and paternal ancestral lines did not cross > and could not be more different in origin, but that is no reason not to > upload both in a single GEDCOM. >
Like everyone else, I have the full complement of matches who don't answer, matches that seem to be more than unlikely and matches that are tantalizingly close, but no cigar. And yes, it is discouraging, but tonight I got a refreshing "gift". One of my matches who lives in Poland e-mailed me and he has found one of his Polish ancestors was from Germany and shares a rather unusual German surname with me. Now not only was this a surprise from a man living in a country I have no ancestors from, but also a surprise that I didn't have to do all or the majority of the footwork. However, I think this is only the second time anything like this has happened. Roberta Estes -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Diana Gale Matthiesen Sent: Monday, May 14, 2012 10:52 PM To: [email protected] Subject: Re: [AUTOSOMAL-DNA] atDNA Survey On my mother's side, I have deep New England colonial ancestry, with dozens of couples in my ancestry who arrived during the "Great Migration" of the 1630s. What probably depresses the number of matches I have is that my father's paternal grandparents were Danish immigrants, from a small family with few descendants (in Denmark or the U.S.). In fact, I just realized... Except for the known cousins I deliberately tested, all of my confirmed FF matches have been with people who also share these well known, prolific New England progenitors. I'm getting just the result one would expect, that is, the most matches where I need them the least. How discouraging. Diana
Diane, Another way to look at it that you're finding the matches on the very well documented lines. Among your atMatches who can't find a Common Ancestor with you, there are probably Common Ancestors with your lesser known lines. All the more reason to include ancestors out to 10th cousins and to add in alternative/iffy/probable/good-guess ancestors. If you still don't find more Common Ancestors, you'll need to start looking hard at matching place/times. atDNA is a great tool, but it takes a lot of work. And some luck.... Good luck, Jim Bartlett On 05/14/12, Diana Gale Matthiesen<[email protected]> wrote: On my mother's side, I have deep New England colonial ancestry, with dozens of couples in my ancestry who arrived during the "Great Migration" of the 1630s. What probably depresses the number of matches I have is that my father's paternal grandparents were Danish immigrants, from a small family with few descendants (in Denmark or the U.S.). In fact, I just realized... Except for the known cousins I deliberately tested, all of my confirmed FF matches have been with people who also share these well known, prolific New England progenitors. I'm getting just the result one would expect, that is, the most matches where I need them the least. How discouraging. Diana > From: Jim Bartlett > Sent: Monday, May 14, 2012 1:35 PM > > Diane - you requested a survey: > > For Family Finder I have 249 matches (probably due to deep > Colonial VA > roots); suggested relationships are 18 are up to 3rd cousins, and > 49 are 4th > cousins. But I can tell from my experience all are somewhat > greater. I know > all of my 16 2G-grandparents; and 30 of my 32 3G-grandparents > (4th cousins); > and very few of my matches are from those generaitons. <snip> ______________________________ For answers to Frequently Asked Questions about mailing lists, please see: [1]http://dgmweb.net/MailingListFAQs.html ------------------------------- To unsubscribe from the list, please send an email to A[2][email protected] with the word 'unsubscribe' without the quotes in the subject and the body of the message References 1. http://dgmweb.net/MailingListFAQs.html 2. mailto:[email protected]
On my mother's side, I have deep New England colonial ancestry, with dozens of couples in my ancestry who arrived during the "Great Migration" of the 1630s. What probably depresses the number of matches I have is that my father's paternal grandparents were Danish immigrants, from a small family with few descendants (in Denmark or the U.S.). In fact, I just realized... Except for the known cousins I deliberately tested, all of my confirmed FF matches have been with people who also share these well known, prolific New England progenitors. I'm getting just the result one would expect, that is, the most matches where I need them the least. How discouraging. Diana > From: Jim Bartlett > Sent: Monday, May 14, 2012 1:35 PM > > Diane - you requested a survey: > > For Family Finder I have 249 matches (probably due to deep > Colonial VA > roots); suggested relationships are 18 are up to 3rd cousins, and > 49 are 4th > cousins. But I can tell from my experience all are somewhat > greater. I know > all of my 16 2G-grandparents; and 30 of my 32 3G-grandparents > (4th cousins); > and very few of my matches are from those generaitons. <snip>
There wasn't much genetic genealogy used this time, but I thought it was still an interesting episode. My review: http://www.yourgeneticgenealogist.com/2012/05/finding-your-roots-with-henry-louis_14.html CeCe www.yourgeneticgenealogist.com www.studiointv.com
Diane - you requested a survey: For Family Finder I have 249 matches (probably due to deep Colonial VA roots); suggested relationships are 18 are up to 3rd cousins, and 49 are 4th cousins. But I can tell from my experience all are somewhat greater. I know all of my 16 2G-grandparents; and 30 of my 32 3G-grandparents (4th cousins); and very few of my matches are from those generaitons. 51 of my atMatches and I have confirmed paper trials to Common Ancestors (7 of them from the old Affy chip). I have an additional 38 atMatches at 23&Me with confirmed paper trails to Common Ancestors (from 1133 atMatches - most still Anonymous, despites various tries to make contact). These 89 Confirmed Ancestors have resulted in several conclusions: 1. Some represtent different ancestors (from one of my parents) on the same segment, proof that at least one of these Common Ancestors is NOT the one provviding the matching atDNA segment. 2. Almost all of these matches are 2-3 generations farther than the (conservative) suggested relationships. Yes, it could be that most of these aren't really matches on the shared segments, but with the number of matches I have and the full Trees both me and some of my atMatches have, I would expect to find a higher percentage (higher than almost zero). Why would so many matches be behind brick walls? 3. We need to pay attention to the segments. I am now mapping the FF and RF segments in one spreadsheet (sorted by Chr and start location, so the segments are in roughly the right order and I can inspect them along each chromosome). Two ancestors from the same parent CANNOT occupy the same place on a Chromosome! Although you CAN have one ancestor from each parent at the same place on a Chromosome - in fact you will have maternal ancestors covering all the Chromosomes AND you will have paternal ancestors also covering all the Chromosomes. Side-by-side smaller segments will usually be from different ancestors. When I talk about segments, I am referring only to large segments. 7.7cM or higher for FF, and 5.0cM or higher for RF (23&Me) - these are the cutoffs the two companies use and every one of your atMatches will have at least one segment of this size. I think the take-away from this is that it's important to carefully look over your atMatches' Trees for multiple Common Ancestors - this is not uncommon with Colonial America roots. Learning a little more every day... Jim Bartlett On 05/14/12, Diana Gale Matthiesen<[email protected]> wrote: I guess it all depends on what you call "a lot," and whether finding "a lot" was your goal. I'll consider the money well spent and the effort worth while if it helps me break through even *one* brick wall. Perhaps we should do a survey... I have a total of 186 matches, over 50 of which are 4th cousin or better, and 10 of which have been confirmed on paper (i.e., a check of their GEDCOM shows our paper pedigrees intersect). Is that a success or a failure? The impediment here is that the majority of subjects have not uploaded a GEDCOM. In fact, I just checked... only 61 out of the 186 have uploaded a GEDCOM. The other hindrance is the number of people who do not include locations in their GEDCOM. For common surnames, I'm simply not going to pursue a lead where I don't know the family is in geographic proximity to mine. Diana
On 5/14/2012 4:40 AM, [email protected] wrote: > I would need separate GEDCOMs for my parents who were from 2 different > parts of the world. Their ancestral paths didn't cross. Therefore, it is > necessary to concentrate heavily on the geography. > > > --Ida Skarson McCormick, [email protected] > > In the begining, FTDNA required separate GEDCOMs for the maternal and the paternal line. This is no longer the case. In fact, it is much easier for all, if one just uploads an Ancestor tree GEDCOM so that both maternal and paternal lines are displayed. Like you, Ida, my maternal and paternal ancestral lines did not cross and could not be more different in origin, but that is no reason not to upload both in a single GEDCOM. After all, we have no idea as a rule when someone with an autosomal match will recognize something in one of the lines as there is no distinction in the match as to whether it is in the maternal or paternal side of one's genome. Such a distinction, called phasing by some, would indeed be wonderful as it would immediately halve most of the possibilities for a match by eliminating one line or the other (except of course when cousins marry cousins). But for most of us, that is not the case. By erecting yet another barrier or excuse towards creating and uploading your ancestral GEDCOM, you are not getting your money's worth out of your atDNA test. These tests are fishing expeditions for the most part and do not work without at least some disclosure on the Internet so that others can inspect your data and ancestors. You take the test and hope that you can connect with someone who knows more than you do. Walter
> From: [email protected] > Sent: Monday, May 14, 2012 4:40 AM > > There are other reasons not yet mentioned for avoiding > GEDCOMs. GEDCOM needs to go the way of the dodo. GEDCOM has been the standard format for exchange of genealogical data for decades, and it will remain the standard. You aren't going to get an entire industry to change to another standard when this one is so thoroughly entrenched. What we can expect is that the standard will occasionally be revised and improved, which it is. > GEDCOMs are inadequate for names from other cultures, > that is, ones which did not have fixed surnames and > ones which require a good deal more space than the > standard American name to identify a person. The GEDCOMs > get very messy. I think it's unrealistic to expect one standard to apply to every possible naming system, not to mention every language. What would a Chinese version of RootsMagic look like, and how would you merge Chinese and European pedigrees? Yes, there is a European bias here. It's unlikely there would ever be a single file format that would be universally, as in globally, applicable. That's no reason for the millions of people using GEDCOM for data exchange to abandon it. > I would need separate GEDCOMs for my parents who > were from 2 different parts of the world. Their > ancestral paths didn't cross. Therefore, it is > necessary to concentrate heavily on the geography. Yes, if the cultural and language differences (especially the alphabet) are that great, you will need to use two different kinds of software to manage the two different databases. I keep my maternal and paternal ancestry in two different databases as it is, just to keep the file size down. > Most people who contact me about American families are > not far enough back in their research for us to tell > where we might have a match. Yes, I'm in the same boat. I suspect most of us are. > If they can zero in on a location (US county) or > other factor that provides clues, I give them > advice to try to bridge the gap and/or explain why > even though the surname's the same, the people > aren't apt to be related. Again, yes, we all face that problem, especially with common surnames. > An ancestor box chart generated with genealogy > software works much better than a GEDCOM for > display purposes. The boxes can expand to fit the > number of words in a name and the amount of space > each person on the chart needs. GEDCOM is only a data exchange format. It has nothing to do with how the data are displayed. How the data are displayed depends on the software that the GEDCOM is *imported into*. If your complaint is how the GEDCOMs are displayed by FTDNA, complain to FTDNA. It's their software doing the display. > A few years after Y testing through FTDNA, my brother > still does not have a close match, most likely due to > lack of fixed surname(s) on the Y-line. Whether or not you have a Y-DNA match has *nothing* to do with anyone's surname. That's the whole point of being Y-DNA tested. If you have a common biological ancestor on your patrilineal line, you will show up as a Y-DNA match *regardless* of your surname or lack of one. > Lack of fixed surnames seems to inhibit testing, > although it should not. I don't see how the lack of a fixed surname would inhibit someone from testing. In fact, it should encourage them because Y-DNA will tell them what their surname cannot. It's the main purpose behind my regional Danish Demes project. I think the only Europeans later than Danes to adopt surnames were the Norwegians. Both cultures had to finally be forced to adopt them, by mandate of law, late in the 19th century. If we Danes are going to connect beyond the 19th Century, it will have to be via DNA. Diana
Hi Everyone, I am new to this list and to the use of autosomal dna but I wanted to share what I have been doing that is turning out to be quite useful. I have created a FileMaker database containing information on most of my autosomal dna matches. I have structured this database so that once I enter the chromosome number, matching segment information, cM and SNP data for each of my matches (when available), I can then view "my dna results" as seen through my matches. I also scan family names from those matches who have provided them and enter what I think might be one or more possible surnames where we might match. I have enough matches (270 in my database so far) that there are some that overlap in both the chromosome and segment. (Gedmatch is another way I check for this.) Even though these matches could come from either my maternal or paternal side, I am getting good responses when I contact these matches and suggest how we might be related. I am also focusing on developing a list of cousin surnames that may not be my direct ancestors but who are descendants of my ancestors. Some of these I am beginning to recognize by scanning "my dna results." I am also creating a genetic profile of one of my most prolific ancestral couples. This also gives me a way to identify some potential matches who haven't provided enough or any family data. I am pleased with some of the responses I've gotten from cousins and have developed some potentially useful leads. Of course there are those who don't respond but this system can give me some useful work-a-rounds for those who don't. There is definitely a lot of information to be analyzed with this autosomal dna testing! I hope I've managed to explain what I am doing in a way that is understandable. If not, feel free to ask me questions. Margaret
I guess it all depends on what you call "a lot," and whether finding "a lot" was your goal. I'll consider the money well spent and the effort worth while if it helps me break through even *one* brick wall. Perhaps we should do a survey... I have a total of 186 matches, over 50 of which are 4th cousin or better, and 10 of which have been confirmed on paper (i.e., a check of their GEDCOM shows our paper pedigrees intersect). Is that a success or a failure? The impediment here is that the majority of subjects have not uploaded a GEDCOM. In fact, I just checked... only 61 out of the 186 have uploaded a GEDCOM. The other hindrance is the number of people who do not include locations in their GEDCOM. For common surnames, I'm simply not going to pursue a lead where I don't know the family is in geographic proximity to mine. Diana > -----Original Message----- > From: [email protected] [mailto:autosomal-dna- > [email protected]] On Behalf Of Kathy Johnston > Sent: Monday, May 14, 2012 1:34 AM > To: [email protected] > Subject: Re: [AUTOSOMAL-DNA] FF clients not uploading GEDCOMs > > > Forgive me, but you are not being at all realistic about > > the odds of finding a relative. > > I think you will find relatives, but you are not likely to find a > lot of 3rd and 4th cousins.? Your matches are much more likely to be > very distant. To be realistic, you really need to do the math. > Figure out how many 3rd and 4th cousins you expect to have living > today on this planet and then figure out the number of people who > would have to be randomly tested to get a large enough sampling of > the population to reasonably expect that a few actual 3rd or 4th > cousins will be among them. The majority of those predicted or > suggested to be 3rd or 4th cousins on your match list simply are not > likely to be 3rd or 4th cousins based on probability alone. You need > a critical number of about a million testees to reasonably expect > even just a few to show up. You can figure 4 children per couple or > even 8 children per couple who lived to have more children and you > would have to agree that FTDNA has certainly not met the critical > number of testees to reach the significant level needed. Ke! > ep in mind there are about 313,000,000 people living in the U.S. > alone and only a tiny percentage of those have been tested. > > Kathy J.
I think it would help if they wrote adopted too so that at least matches know that is the reason no distance ancestor name is written. There seems to be so many unfilled in distant ancestors amongst my matches that they can't all be adopted. If people sign up to do the test it would really help if they filled in the forms to help themselves and others. Karen On Mon, May 14, 2012 at 1:28 AM, John F Smeltzer <[email protected]>wrote: > > > > > > In the FTDNA system at least if one simply adds "Adopted" in the surnames > listings .... it helps a LOT.. Again ... a suggestion to be added to a > list that we could collectively craft here on this list and provide to > FTDNA as a "How to" suggestion. Alternatively .... we could send it out > ourselves. Such a "How to" would be very useful to many. > > > > John > > > > :::::::::::::::::::::::::::::::::::::::::::::::: > > I agree Mike .... I am assisting several folks who are either adopted or > have an adoption in their line that has been their biggest brick-wall. > And many of them eagerly (and with great excitement) paid their money, took > the test and now are uncertain (or are at least struggling) as to what to > do. I can understand the enthusiasm that may have brought them into the > game. When I run into someone with significant matches to our group of > kits I always try to include them as collaborators if they are > willing. Including adoptees ... My theory is that eventually we will have > enough pieces to the puzzle to at least be able to offer deep ancestral > story lines even if we can't provide near-term names. In the meantime I > find that by working in collaboration with these folks who show interest I > learn a lot and that helps build my understanding of how best to apply > these tests in the real world. It's not without its faults but it seems > to be at least one successful and rewarding strategy. > > And, I think we shouldn't jump past the fact that w e are all still > learning how to best apply these Family Finder tools. There is no one > singular cookbook way to do this. And while I agree that GEDCOM's and / > or Pedigrees are great ..... a simple (but comprehensive) surname list is a > great and often sufficient start. My suggestions to struggling test > participants are always step-wise ..... 1. develop a complete surnames > list and enter it into FTDNA (or other vendors) files as they allow. Then > ... as time permits 2. Build your pedigree ..... stretching your > genealogically imposed need to proof everything to its limits .... just > don't declare it all to be fact and don't post it on-line as fact. And > .... in those spare moments of time ..... 3. Build out your family files > as far as you can along as many lines as you can ... with particular > emphasis on working out sibling lines of more ancestral family to include > spouses and at least a generation or two downstream ... which allows you to > pick up on other potentially matching surnames from present day testing. > Surnames that are NOT in your direct line but ones that IF (and when) > TESTED will in all probability match your targeted surname group. > > Now, that said .... even in my own situation I have so many kits (both > recruited cousins and collaborators) in the files ... over 50 at present > ..... it's difficult at best to work matches AND do upgrades while trying > to conduct normal day-to-day business at the same time. So, as in most > everything we do these days ... most of us are behind on more things than > we would like to admit. Now, when someone contacts me and engages in a > polite conversation about our matches I can re-direct if necessary to help > fill in those GEDCOM / Pedigree / surname gaps if they exist and if I can > do anything to accomodate. But, I find it almost as interesting that > even with a LOT of test kits out there that the number of "incoming" > contacts is much much fewer than I ever anticipated. It's my "out-going" > communications that generate most of my activity. And I believe this is > the case because MOST folks just simply don't know what to do ... where to > go .... who to turn to and have the normal human emotion of not wanting to > look badly in public / or public communications. I've never been > hindered by that burden. I can look "stupid" most anywhere and be ok > with it as long as I'm trying. > > So .... I think a one page outline of "best practices" .... would be very > useful. It could be distributed as a "cheat-sheet" to folks who have been > waiting to get a sense of what to do next. And, we could provide it > to our matches as a pro-active gesture that shares our experiences with > them in a way that might just prod them to those so needed next steps. > > This is complicated stuff .... the more we can help uncomplicate it for > the masses the more productive our own efforts will be .... > > John > > > > > > ______________________________ > For answers to Frequently Asked Questions about mailing lists, please see: > http://dgmweb.net/MailingListFAQs.html > > > ------------------------------- > To unsubscribe from the list, please send an email to > [email protected] with the word 'unsubscribe' without > the quotes in the subject and the body of the message > > > ______________________________ > For answers to Frequently Asked Questions about mailing lists, please see: > http://dgmweb.net/MailingListFAQs.html > > > ------------------------------- > To unsubscribe from the list, please send an email to > [email protected] with the word 'unsubscribe' without > the quotes in the subject and the body of the message >
Kathy - take the plunge - you will be rewarded for it. atDNA is a very poor tool to test a theory - there are just too many variables. It's a great tool to find cousins and "prove" some blood lines. I've put my comments in brackets by your excuses... Jim Bartlett On 05/13/12, Kathy Johnston<[email protected]> wrote: OK. Here are my lame excuses to add to those already mentioned. There are several problems in producing a? GEDCOM: 1. My surname list is too big and I will need to retire from my job before I have time to generate a decent GEDCOM. It will go on forever. [As Diane points out, the GEDcom needed for FTDNA is finite and probably what you have is within range] 2. I don't want to publish something that has any potential mistakes because those mistakes will get repeated over and over. It takes too much time to verify it all. [for the purposes of atDNA matching, don't worry about mistakes - your atDNA matches are scanning your list looking for ONE common ancestor (actually a husband and wife) - they don't care about all the others] 3. The farther you go back in time, the more likely there are questionable ancestors and multiple matching lines. [yes, and what better way to sort them out than by finding other cousins who have researched the same lines] 4. I want people look at my surname list with geographical locations next to each name and not just a tree. Some of the surnames on my list are possible or probable but not proven. Maybe there is a middle name that sounds like it came from an ancestor and I believe I am on the right track because clearly the name, date and location show that surname in the same place at the same time.? Getting over those brick walls may mean showing information that is not yet proven in the pedigree. [for the atDNA matching purpose, it doesn't have to be proven - again, your atMatch is looking for ONE common ancestor, and will gloss over all the rest] 5. Even if I go back 4 generations, am I likely to randomly match any true 3rd cousins? NO!! because you need a database of over a million to find that one in a million 3rd cousin. Even if you have more than 300 3rd cousins alive today, there are over 300 million people living in the U.S. Does FTDNA test that many people? No. Are the 3rd and 4th cousin predictions accurate? Not at all because in general you are picking up the more distant cousins right now that are falling out of their expected cM range. Some of these segments are probably over 2 standard deviations above their expected size so your chance of finding the ancestor is low. What I am saying is that what looks like a 3rd or 4th cousin is really a very distant cousin. The database has not yet reached a critical sample size. So is it really worth all the effort right now? [Look at it from a different direction: all of your atMatches share a large segment of atDNA with you, and are highly probable cousins! Take that at face value and try to determine how your relate. IMO, this is a genealogists dream - getting a list of new probable cousins with whom to share] 6. Family Finder (or Relative Finder) works best if you are trying to test a hypothesis. ?For example, on the television show, at PBS, "Finding Your Roots", Henry Louis Gates, Jr. used autosomal DNA to see if an African American could have been descended from a slave owner by testing a presumed 3rd cousin who had a good paper trail. That is how this autosomal DNA should be used successfully, as a tool to test a hypothesis. These tests are not nearly as good at finding those random matches that everyone is talking about. [I disagree! atDNA is poor at testing a hypothesis, just because the shared segment may not be there for folks who are actually cousins. I've found over 60 new cousins from my atMatches - they may not all be on the shared segment, but that's what I'm finding out now through Phase III: segment mapping] So that is why I am in no hurry; I admittedly do need to work on getting a GEDCOM updated, but maybe for other reasons... [Take your time if you want, as for me, I'm trying to get as much out of atMatches as I can (we've already lost a lot of 23&Me matches who dropped out after one year. Through segment matching, I know some of their ancestors now, but can't contact them...] Kathy J.
Thanks for the kind words, Jim. As a result of feedback to my prior post on the topic sent to me by private email, and subsequent reading on the subject, I am going to revise my method. The issue is outgoing messages never making it to the match's inbox because it gets labeled as unwanted in advance by the email clients or other structures on the route. I will not go into examples at great length here because the interested reader will be able to search and find better articulated treatments of the topic elsewhere. But I will mention a few, just to get the idea across. The idea of capitalization. Orthodoxy in genealogy has historically been to put surnames in caps, so that people can find them quickly in text. And that is great. The problem is that all caps in a message triggers the filters. That is especially true for the subject line. So the autosomal that appears in the subject line of this message, while serving its purpose, is probably net negative for filtering. Use only plain text. Higher order, that is to say formats that go beyond plain text, have a tendency to set off filters. So I am still going to send my message out the same way, but I am going to convert it first to plain text first. I noted my surnames tables in the archives of this list looked okay, so I believe it will work. And in retrospect, there wasn't much gained by the shading of the text. Certain non-plain text elements are likely to trigger filter. These are tags that get interpreted by the email client as pictures or links. The email clients use them as code. The filters see them as bad. So that is another reason not to use them. Keywords should be avoided. Certain words just send the filters into a frenzy. You surely know what they are, because you see their unwanted selves in your messages. But if you don't, look online for worst words to send, and avoid those where possible in your emails. Abbreviations and certain like jargon should be avoided. If it is not a normal word to the filter, it should be avoided. This is because people who want their unwanted email to get through the filter will break up keywords with punctuation and other devices. The filters are aware of this, and if your text looks like that, even while innocent, then it may get tagged. Sender domain should have a good reputation. I have used yahoo mail for my messages. I would be better served by using my other email account I get from my cable company, which I have never used thus far. If I am sending out many messages at once from a yahoo account, the filters do not like that. Hopefully with these revisions I will get a better response rate. Gregg --- On Sun, 5/13/12, Jim Bartlett <[email protected]> wrote: From: Jim Bartlett <[email protected]> Subject: Re: [AUTOSOMAL-DNA] Improving FF match email response rates To: "Gregg Bonner" <[email protected]>, "[email protected]" <[email protected]> Date: Sunday, May 13, 2012, 10:57 PM Gregg A very well thought out process, designed to make it easy for our atMatches to easily recognize our Common Ancestor. I use a similar process and have about a 75 percent (eventual) response rate with Family Finder, but only about 30 percent with Relative Finder. By far, the FF matches are more serious about genealogy. I like to include a Patriarch if I feel that most genealogists will recognize him - the aha! factor is very powerful - if someone sees their ancestor on my list, they will respond quickly. I think it is imperative to include at least as many of your 512 ancestors as you can. Between FF & RF I have over 1,300 matches. I'm using those to map segments on my chromosome. I estimate about 200-250 segments per parent, or 400-500 total. Although some ancestors will provide more than 1 segment, it seems many of the 512-level ancestors (8th cousins) will show up. If you have some sticky segments that tend to last for extra generations, it could come from a generation or two further back. As a further check of this, you probably have about 1600cM from each parent. Dividing by 7.7cM (for the FF cutoff) you get about 200 such segments, or 400 for both parents. You'd need at least some 8th cousins for these segments. Thanks again for outlining a great process. Jim
There are other reasons not yet mentioned for avoiding GEDCOMs. GEDCOM needs to go the way of the dodo. GEDCOMs are inadequate for names from other cultures, that is, ones which did not have fixed surnames and ones which require a good deal more space than the standard American name to identify a person. The GEDCOMs get very messy. I would need separate GEDCOMs for my parents who were from 2 different parts of the world. Their ancestral paths didn't cross. Therefore, it is necessary to concentrate heavily on the geography. Most people who contact me about American families are not far enough back in their research for us to tell where we might have a match. If they can zero in on a location (US county) or other factor that provides clues, I give them advice to try to bridge the gap and/or explain why even though the surname's the same, the people aren't apt to be related. An ancestor box chart generated with genealogy software works much better than a GEDCOM for display purposes. The boxes can expand to fit the number of words in a name and the amount of space each person on the chart needs. A few years after Y testing through FTDNA, my brother still does not have a close match, most likely due to lack of fixed surname(s) on the Y-line. Lack of fixed surnames seems to inhibit testing, although it should not. Ironically with my own testing on 23andMe, it is the closer matches who fail to respond. --Ida Skarson McCormick, [email protected]
> Forgive me, but you are not being at all realistic about > the odds of finding a relative. I think you will find relatives, but you are not likely to find a lot of 3rd and 4th cousins.? Your matches are much more likely to be very distant. To be realistic, you really need to do the math. Figure out how many 3rd and 4th cousins you expect to have living today on this planet and then figure out the number of people who would have to be randomly tested to get a large enough sampling of the population to reasonably expect that a few actual 3rd or 4th cousins will be among them. The majority of those predicted or suggested to be 3rd or 4th cousins on your match list simply are not likely to be 3rd or 4th cousins based on probability alone. You need a critical number of about a million testees to reasonably expect even just a few to show up. You can figure 4 children per couple or even 8 children per couple who lived to have more children and you would have to agree that FTDNA has certainly not met the critical number of testees to reach the significant level needed. Keep in mind there are about 313,000,000 people living in the U.S. alone and only a tiny percentage of those have been tested. Kathy J.
Gregg A very well thought out process, designed to make it easy for our atMatches to easily recognize our Common Ancestor. I use a similar process and have about a 75 percent (eventual) response rate with Family Finder, but only about 30 percent with Relative Finder. By far, the FF matches are more serious about genealogy. I like to include a Patriarch if I feel that most genealogists will recognize him - the aha! factor is very powerful - if someone sees their ancestor on my list, they will respond quickly. I think it is imperative to include at least as many of your 512 ancestors as you can. Between FF & RF I have over 1,300 matches. I'm using those to map segments on my chromosome. I estimate about 200-250 segments per parent, or 400-500 total. Although some ancestors will provide more than 1 segment, it seems many of the 512-level ancestors (8th cousins) will show up. If you have some sticky segments that tend to last for extra generations, it could come from a generation or two further back. As a further check of this, you probably have about 1600cM from each parent. Dividing by 7.7cM (for the FF cutoff) you get about 200 such segments, or 400 for both parents. You'd need at least some 8th cousins for these segments. Thanks again for outlining a great process. Jim - Sent from my iPhone - FaceTime! On May 13, 2012, at 4:16 PM, Gregg Bonner <[email protected]> wrote: > Dear Listmates, > > I have done some research on this, which was mostly in the form of asking for a copy of people's intro email (for those whom I contacted first), and asking them their email response rates. Then with a little subjective correlation analysis, I came up with the following guidelines: > > 1. Make your introductory email VERY succinct. > > I know you or I might read a tome for each and every match, but most people won't. The worst thing you can do is give an email which immediately evokes the "too long; didn't read" response. I recommend something along the lines of the following for the introduction: > > "Dear $name_of_match, > > FTDNA tells me that $name_of_match and I share significant stretches of DNA. I'd like to exchange lists of ancestral surnames to see if we can find our common ancestor from whom these DNA stretches originated." > > (I use the actual name of match instead of "you" because people may have several proxy samples they administrate under the same email address, and you may not match them all...so that may save a round of emails that would otherwise call for a disambiguation of who the match is to - but I digress....) > > 2. Include your list of ancestral surnames right there in the intro email. Do not make anyone click any link. Do not make them hunt down your public tree on ancestry. > 3. Make your list of ancestral surnames alphabetical. They know what they are looking for. Make it easy to find. > 4. Make your list of ancestral surnames short, to the point where it is realistic to believe that any shared DNA could have come from the surname on the list. > > The reader should be able to look at the list and determine if he has any interest in the surnames on the list more or less instantly. I know people will say that they emailed their match, and by comparing surnames, they found their common ancestor who was born in 1530. Probably not. You probably just didn't find the more recent ancestor who was the real origin of that shared DNA. By making a mile-long list of surnames, you are probably making it less likely that they will read them. > > So, for my part, I include (paste a copy of) an HTML table that gives just 32 surnames (in my cases, I tested my mom and my dad, so it is 32 ancestral surnames for each of them - I just give the 32 for which parent of mine they matched in the intro email). That number of ancestors from a parent takes me to 5th cousins, which I think is about the limit of the test's resolving power. > > The table is two columns, with the alpha surname list left, and on the right the migration path. The considerations for the rightward column are these: > > A. Adding the number of characters to the number of characters in the surname should not exceed 72. I don't want the text to wrap to the next line and mess up the register...for cases where their email client isn't set up to handle HTML tables well (similar to this rootsweb list). > B. Since you won't have enough characters to describe the migration path completely, it should concentrate on where that surname was in the timeframe from about early 1700s to 1850s or so. Earlier than 1700, and the test isn't really picking that up as a match, and later than 1850 or so, then you are close enough cousin that your match should "scream out", and you'd probably know who the person was anyway. > C. Pack as much information as you can in the 60-something character spaces you have for migration path. > > I think it is important to NOT put a patriarch's name. If you list George Surname, and that is one generation above their Fred Surname, and they don't know about George, then they could think, "well, mine's not George, so I'm not interested". A matching surname and the right place at the right time will get a response if anything will. > > 5. The only thing I add to the very brief intro (above), and the 32-row-long two-column table that follows is the plea for them to respond with their own list of ancestral surnames, articulating the rationale that they might not see a connection, but you could very well recognize their surname as an important collateral line to yours, and thus extend their ancestral line, even though they would not have noticed it. > > My response rate is inexplicably low, but it seems still better than most. I have a 25% response rate that is almost immediate, and that improves to about 40% with a 2nd email (after a month or two, I send out a "I think maybe your spam filter got my first email" message). There has not been enough time for me to do a 3rd yet, and eventually I will send out a weekly email to non-bouncing addresses for 10 to 15 consecutive weeks (and then write those off and remove them from the email list). I suspect that my total response rate will approach 50%. > > Gregg Bonner > P.S. To see the surname tables I describe in their intended HTML format, see the 2nd and 3rd table at this URL: > http://freepages.genealogy.rootsweb.ancestry.com/~gbonner/DNA/FF/email.html > > P.P.S. To see this tabular information in this message, in a way that will most likely lose format, and get mangled beyond all readability, see below (in my emails, the surnames are shaded according to genetic closeness - to see what I mean, click the link above): > > [to dad's matches] > > D. W. Bonner (FTDNA# 209176) > shade code = 1st cousin 2nd cousin 3rd cousin 4th cousin 5th cousin > [--?--] Chatham, NC late 1700s - 1830s > Aiken SC; GA; AL > Armstrong Augusta, VA 1710s; Mecklenburg, NC 1760s; Bedford Co., TN 1840s > Bonner Pr Geo, VA 1700s; Troup, GA 1830s; Chambers, AL 1850s; OK > Brigance PA, Sumner Co., TN 1810s; Bedford Co.,TN > Brown Meriwether, GA > Buchanan Augusta, VA 1730s-1770s; Sumner, TN 1810s; White, IL 1840s > Bussey Calvert, MD 1730s; SC 1790s; Bedford, TN 1800s > Carlisle IRE 1700s; SC late 1700s; AL 1800s > Chappell VA 1700s; Troup, GA 1800s > Connelly Chester, SC 1790s; Troup, GA 1830s; Randolph, AL 1890s > Cottle NC 1800s; Monroe, GA 1830s; Chambers, AL 1860s > Formby VA; Troup, GA > Gant Bute, NC 1750s; Abbeville, SC 1800s; Bedford, TN 1850s > Grant Abbeville, SC 1790s; Randolph Co., AL 1840s > Hamble Bedford, TN early 1800s > Jones #1 Troup, GA 1830s; Chambers, AL 1840s; OK 1930s > Jones #2 Guilford, NC 1770s; Sumner, TN 1800s > Jones #3 GA; AL; (AR) > Lemaster SC 1700s; AL 1800s > Lentz Fairfield, SC 1750s; Rowan, NC 1780s; Bedford, TN 1850s > Mayfield Warren/Bute, NC 1760s; Granville Co., NC 1780s > Moore VA 1700s; Troup, GA 1800s > Neeley NC; Bedford, TN > Pickle Orange, NC 1780s; Bedford, TN 1850s; Marshall, TN 1900s > Rutledge Georgia or Alabama 1800s > Shearin Bute, NC; Bedford, TN > Sikes Halifax, VA 1760s; Lunenburg; GA 1790s; Bedford/Rutherford, TN > Springer SC 1700s; Bedford, TN 1800s > Stephenson Amherst, VA 1700s; Bedford, TN 1800s > Williams #1 NC 1780s; Sumner, TN 1810s; Bedford, TN 1850s; OK 1930s > Willingham Lunenburg, VA 1700s; Columbia, GA 1810s; Randolph, AL 1860s > > === > > [to mom's (Happy Mother's Day, Mom!) matches] > > Ann Simpson (FTDNA# 227381) > shade code = 1st cousin 2nd cousin 3rd cousin 4th cousin 5th cousin > [--?--] probably DE > Apperson VA 1600s-1700s; Davidson, TN 1800s > Barrett (Baltimore?) MD 1760s; Ohio Co., KY 1810s-1860s > Buck Schleswig-Holstein > Classen Schleswig-Holstein > Coleman Cumberland/Buckingham, VA 1750s; Davidson, TN; Logan, KY > Dutch Franklin Co., PA 1790s-1810s; Preble Co., OH 1830s > Fleagle Bucks, PA 1740s, Frederick, MD 1770s; Preble, OH 1850s > Gentry VA?; KY?; Gibson, IN > Haack Schleswig-Holstein 1600s-1890s; IA; OK > Harder Schleswig-Holstein > Harris Rutherford, TN > Hartwig Schleswig-Holstein > Jones #4 Rutherford/Davidson Cos., TN 1810s-1860s; St. Louis 1830s > Jones #5 Ohio, KY > Keith Hardin Co., KY 1820s; Ohio Co., KY 1860s > Knowles Sussex, DE 1750s; Greene, GA 1810s; Gibson, IN 1850s > Kraage Schleswig-Holstein > Maassen Schleswig-Holstein > Mills Frederick Co, VA 1740s; Guilford, NC 1770s; Jay Co., IN > Phelps Rutherford/Davidson Cos., TN 1810s-1860s; St. Louis 1830s > Ramey France 1600s; Westmoreland, VA 1700s; Frederick, VA 1800s > Reed Greene, GA 1800s; Gibson, IN > Schlosser GER; PA; OH? > Smitz Adams Co.(?), PA 1800s; Lancaster, PA; Preble Co., OH 1860s > Smotherman ENG; VA; Rutherford, TN; OK > Spore Frederick, VA; Gibson, IN; OK > Swallow Kent Co., DE 1740s; Stokes, NC 1780s; Montgomery, OH > Waack Schleswig-Holstein > West MD; Rutherford, TN? > Wheeler Sussex, DE 1730s; Kent, DE; NC 1790s > Williams #2 Brunswick Co., VA 1790s, Rutherford Co., TN 1840s > > The right-hand side evolves to improve as I find better ways to cram information into 60-something character spaces. > > So much for being succinct.... > > > > ______________________________ > For answers to Frequently Asked Questions about mailing lists, please see: > http://dgmweb.net/MailingListFAQs.html > > > ------------------------------- > To unsubscribe from the list, please send an email to [email protected] with the word 'unsubscribe' without the quotes in the subject and the body of the message
I concur with Diane's suggestions with one exception. I advocate including iffy ancestors in a list of ancestors for atDNA purposes. When we compare ancestors with our atDNA matches it's not like we are promoting our list as correct; we are providing a list of potential matches for our atMatches to review. Our atMatches are searching thru our long list of ancestors for the ONE that we have in common - they have no interest in any of the others. And if they match on an iffy ancestor of ours, then we REALLY have something to talk about. I had a HIGGINBOTHAM ancestor on my list which has been proved wrong by Y-DNA, but one of my atMatches also knew of that ancestor, and we exchanged info on that line. We've discovered that our ancestors were siblings, but the father is unknown. I now know what atDNA segment we share, and can help others who match me on that segment. I say throw in the kitchen sink; add all the possibilities you're working on and hope one of your atMatches has the same possibility in their Tree - you may learn something new; I did. Jim - Sent from my iPhone - FaceTime! On May 13, 2012, at 4:51 PM, "Diana Gale Matthiesen" <[email protected]> wrote: > When you make an "abbreviated" GEDCOM for upload to FTDNA, you should > *NOT* include any iffy connections. Only include the solid > connections. >
Roberta That's what I did - it took some work, but it's been very helpful. Also I took that Ancestry tree and made a GEDcom from it, which I posted on the free World Connect site, and can use at FTDNA. Jim - Sent from my iPhone - FaceTime! On May 13, 2012, at 11:52 AM, "Roberta Estes" <[email protected]> wrote: > I certainly do think it would be easier to start over and make a skeleton > one than to try to modify the existing one. > > Roberta > > -----Original Message----- > From: [email protected] > [mailto:[email protected]] On Behalf Of Diana Gale > Matthiesen > Sent: Sunday, May 13, 2012 11:35 AM > To: [email protected] > Subject: Re: [AUTOSOMAL-DNA] FF clients not uploading GEDCOMs > > Well, you could do what I did... > > I made a new version of my genealogy database, expressly for use in creating > a GEDCOM to upload to FTDNA. It contains only the bare genealogical > information: name, birth date and place, and death date and place. No > sources, no media, no siblings, nothing else, just the pedigree (ancestral > lines). It only goes back 11 generations (12 if you count the test subject) > because that's all FTDNA will display, no matter how many generations you > include in the file. The GEDCOM is only 106K in size. > > Yes, it was a nuisance to create, but IMO, it was worth the effort because I > think it's *that* important to have a GEDCOM uploaded to your FTDNA account. > > > And yes, when I add a new near ancestor (within 11 generations), I have to > update both genealogy files (my regular one and the abbreviated FTDNA one), > then re-export a new GEDCOM to re-upload to FTDNA. But doing so only takes > minutes; and, again, I think it's well worth the effort. > > Diana > > > > > ______________________________ > For answers to Frequently Asked Questions about mailing lists, please see: > http://dgmweb.net/MailingListFAQs.html > > > ------------------------------- > To unsubscribe from the list, please send an email to [email protected] with the word 'unsubscribe' without the quotes in the subject and the body of the message