Another great conference in the books! There is a sale going on and you will find the information at the end of this blog post. Until next year!
My notes . . .
Katherine Borges, Director of ISOGG, welcomed everyone to the ISOGG meeting. Katherine explained why this is called the ISOGG FTDNA Chapter. At an early conference, a man came from another company and tried to solicit customers for his own company. Katherine also shared that she has never had a paycheck from any DNA company. She pays for the ISOGG website. She doesn’t appreciate getting beat up for things like that on Facebook. In the old days, DNA project administrators were discriminated against and were not able to post about their projects on Rootsweb. She was kicked off the Rootsweb Fuller mailing list. Katherine said that it is to our benefit to talk about different DNA companies. It’s what keeps the price low and drives private development. The only thing she doesn’t want to see is bashing of any companies. Take it to the company. Katherine explained that grandstanding is when you bash on a company in a public forum, especially when you don’t take it to them. She noted that she needs to add more moderators on the ISOGG Facebook page who are in different time zones.
The International Society of Genetic Genealogy was founded in 2005 after the first FTDNA conference in 2004. Katherine told the story of how she became ranked right below “expert” at the first conference because she had shown a DNA video to her Colonial James and DAR groups. She reminded everyone that you need to know a lot but you don’t need to know it all. ISOGG is self-funded. Katherine pays for the website and other people pay for other things. People already have to pay a lot of dues and fees. Katherine said that she has been called Benevolent Dictator.
On the ISOGG homepage, it shows that the ISOGG wiki has won the Family Tree Magazine 101 Best Websites award. She showed the wiki welcome page and mentioned that Who Do You Think You Are? Live conference will not be held any more. The word is that another large family history organization in England may start a new conference. FTDNA has been going to WDYTYA since 2009 in London and then in Birmingham.
There is an ISOGG DNA Project Admins Yahoo Group. It’s a private group. It is not ok to share information you read on the list or copy it to other places without permission. There are some company people in there as well as a few people who don’t have projects but are very knowledgeable.
Next, she talked about ISOGG Wiki DNA Interest Groups. If you don’t see a DIG that’s in your area, consider starting one.
When ISOGG goes to England and Ireland, they bring a Sponsored Tests poster. You don’t have to sponsor these tests yourself. You can ask your project members. They can also do it in memory of someone. Don’t miss out on having your surname of interest listed on the poster.
Katherine read an update from Ray banks. Ray would like help from anyone willing and able, particularly in Haplogroup R. Katherine mentioned that people want the tree to match their agenda. Alice set the tree up and Ray maintains it the way they want to. The ISOGG Y-Tree & JoGG have been cited in 206 papers.
There was a new JoGG editor named Leah Larkin. Linda Magellan does the IT and the layout of papers. She could use help with that. Some others are needed for associate editors, articles, paper layout, and peer reviewers. If you are interested, Leah’s email is JOGG@ISOGG.ORG.
Katherine then recognized Derrell, who organized all of the traveling for booths to other countries. She will be retiring.
Katherine then called up Peter Sjolund. The Swedes love to DNA test. Peter had 740 people in the audience for his speech this year and then they all got in line to test. Most people in Sweden speak English. Sometime’s it’s difficult to communicate in different places but it’s very helpful to have such helpful admins. Katherine introduced the four admins. This booth is so busy that they can’t leave for a moment. The Swedish people say they want to match their American cousins. Peter is the ISOGG Regional Coordinator for Scandanavia.
When Peter tested, there were only 300 in Sweden in 2011. Now there are 40,000 people tested.
Next year Genealogy Days will be in Smaland.In total, 25% of the Swedish population went to US. They will be in the heart of Swedish emigration to America.
This year’s record was 500 tested but in Smaland the people don’t like to spend money. They will need donations. Please consider a donation to the Swedish DNA Project General Fund. Donations made during this year will be used in the event in Smaland. They might try something like if you have all 8 great-grandparents in Smaland, you can get a free test.
Ingvar Kamprad went from Elmtaryd to Agunnaryd.
Next Katherine called up Maurice to talk about ISOGG Ireland 2017. They went to Trinity and University College Dublin’s DNA lab. Gerard Corcoran arranges the day. They have a lot of fun. Please feel free to join along. For Belfast in February, Maurice has said he will do Game of Thrones. She thaned FTDNA for being supportive of ISOGG since Day 1.
Maurice thanked both Katherine and Derek. They have been going to Genetic Genealogy Ireland for five years, piggybacked on Back to Our Past, on the Active Over 50 Show. The DNA side of things is really taking off. They had the largest audiences ever at the DNA lectures this year. There were about 150 people in every lecture. The speakers are a mix of our own people. They try to bring up new speakers. This year there were people who had never spoken before. They are balanced by academics. This year there were professors who got funding and traveled to Dublin. The videos will be on Youtube. There is also a blog for Genetic Genealogy Ireland.
This year they set up the ISOGG Ireland group with 14 project administrators who are working in Ireland. There is a new ISOGG Ireland page on the wiki. They will do shows and events around the country and they are being invited to speak at Clan rallies. There are also Irish academic projects such as DNA Atlas.
Legislation in regards to adoptees in Ireland is also being changed. They produced a 27 page document to submit to the government. It will help Irish adoptees in the next 3-5 years. They trained 45 social workers who work with adoption agencies in Ireland and now they are getting referrals. Maurice is currently working with nine adoptees in Ireland to help them find their families. The more the collaborate, the better the outcome for Irish adoptees.
Back to Our Past is going to Belfast February 16-17, 2018. They are hoping to test some of the main actors of Game of Thrones and visit the set. It is held at a conference center called Titanic Belfast.
Cynthia Wells was remembered. Cynthia was a beloved member of the ISOGG group who travels to the various conferences.
The next speaker was Matt Dexter, presenting his success story “Finding His Father—An Adoptee’s DNA Experience.” In 2009, Matt started into the world of genetic genealogy because he couldn’t figure out who his birth father was and neither could his mother. When he was in school he struggled about what to put on his family tree and he listed his adoptive parents.
Matt found out he was adopted when he was about nine years old. The doctors wanted the medical history and his mom said there was none because he’s adopted. As he got older, it started to effect him. He wondered why someone would give up their child. What he really needed was one simple thing—closure. He had good parents, a stable home, and grew up in a great place but there were things nagging him in the back of his mind. He needed his medical records but he wanted to find his father for closure and to answer his unanswered questions. He didn’t know anything about his family except his name and that he had a full sibling, with the same father and mother. His mother wanted to give her to the same family but then changed her mind. He also heard he had Native American ancestry. As he became a teenager, it became more important to him to find his sister than his mother. He felt like Leah and Luke from Star Wars. He wanted to find his secret sister and sneak up and say that’s my sister. His sister Debrahad gone to the state of California and filled a waiver of confidentiality. Fourteen years later he found it. She filed in August 1996 and Matt found the waiver in August 2009.
When Matt started a new job, a co-worker had Native American ancestry and he sent for his non-ID and that’s where he found the waiver of confidentiality. He filed his own and a couple weeks later he got a call from a woman in tears. It was an elderly woman and she said she was his mother. He did a background check on her and it wasn’t good. He found out that his mother had grown up in foster home care and care by other relatives of her parents and she never grew up in a home with her parents. That was a recipe for disaster for this young woman. She was looking for love in all the wrong places in her 20s and Matt was born. He found out that he was given the name of the father. Then he and Debra found out they might not be full siblings. His mom later went on and got married and had a child and the man adopted the other two children. They were told they had a brother who was adopted but Matt never knew he had three siblings.
Matt started building a huge family tree. He didn’t see any Native American. It turns out he doesn’t have any. The other thing he was told were things about his father. He wants to point out that you don’t always want to believe the paperwork. His mom put down Debra’s father’s information as his father because she didn’t know what else to put. His mom said his dad was a man with blonde hair and a certain height. Debra’s father told him who the man he was describing was.
Matt called and told him he was doing research on Pacific State Hospital and asked if he worked there in the 1960s. He said he was looking into information about another employee named Bonita Bailey. He recognized who she was and asked if he was Bobby. He said he’d call back that night with his family and he might be his father. Matt asked if he’d like to know for sure and they did a CODIS STR based test. It takes 13-15 markers and compares it. You have to match every marker to be a parent. The test showed that he wasn’t the father. He also tested Bruce and Debra. He wasn’t her father either. Debra said she didn’t believe the DNA test. For seven years she insisted the DNA test was wrong. Now both of them had no idea who their father’s were but they were also shown as half siblings so they were looking for two different men.
At that point Matt began to look into genetic genealogy and found Family Tree DNA. He had heard of the Y-DNA test so he took it. His closest match at 37 was a GD of 5 with the name “Lucas.” There was a Lucas man who worked with his mother so he found that person and one of his nephews had sons who were the right age and asked them to test to see if they were related. Matt then learned about haplogroups. His was R-M269 and the Lucas people were I-P37.
Matt then learned about Family Finder and began to learn about autosomal. He tested his family, his children and grandchildren, his daughter-in-law, and her parents.
Matt explained about excluded vs. non-excluded.
Morgan is compared to Ron. Ron and Morgan share 3384cM. Ron is NOT EXCLUDED as being the father.
(Ron could have a twin brother named Don and Don is the father.)
We have to use probability. What is most likely or less likely. Most probable vs. least probable.
Matt reviewed how different children inherit differently from each parent and showed this in the chromosome browser.
Matt shared how he reviewed his matches and looked for connections. He became stuck because he didn’t have enough information to check the children of a couple he knew were probably his ancestors. Then he got a match that was a second cousin but she had no family tree. He contacted her and she didn’t seem to know or care who her relatives were. He went through this for a few years. People would match his mother but not have or give information. Through this, he was able to get pictures, things that his mom didn’t have. He found out his grandmother had passed away. This search was beneficial but it didn’t help him find his father.
Then he got a 183cM match but that person was adopted. It was shared with his Irvin cluster. It became tiring and his wife was getting tired of it as well. He was doing it all the time. Then he remembered seeing an Irvin Y-DNA match so he went and looked and it was a GD 2 so he thought it was too far and he was not an Irvin.
Another man matched who shared with the Ervin and with the adopted person. It took a few days to get in touch but the man had found his half-brother. He remembered his dad taking a woman to the hospital and she had a child and the father never saw him again. Matt asked him who the grandparents were and he said his daughter had that info. Matt looked up the grandparents. It turns out his grandmother was an Irvin and the same relatives were in her tree. He had just found one generation closer. That made him either 3rd or 2nd cousin. That was 2016.
Matt began looking further and realized he was looking too close. He backed up a little and saw a Coon connection. He had no Y-DNA matches that were a pattern. Throughout this all, he saw names pop up and tried to figure out what was going on. He saw a match to Henry’s and to Norris people. His cousin Michelle wanted to know, too. She worked hard to figure out how they were connected. She tested a ton of her family. He also matched her first cousin, which narrowed it to one branch. Her half uncle also matched Matt and them at the same place. Michelle was really helpful in finding out the connection through Stephenson. Norris married Sanders, whose mother was a Stephenson. Those parents are the same relatives of Michelle, making them fourth cousins. He began to look at the children of Arthur Henry and Maggie Norris. They used obituaries to figure out that they had 49 and 25-year-old sons when he was conceived.
He called. He had a script ready. “Hello. I’m doing some family research and I noticed that your family and me are somehow related. Have you ever heard of DNA tests for relatives and ethnicity?” No one answer. He called again. No one answered. He hung up again. Later someone called back from that number and someone asked, “Why are you calling me?” He finally explained that he had done DNA testing and matched people in Ohio who were his relatives. He said he didn’t know much about the Ohio family and he left when he was 17. He figured it out. He asked Matt who his mother was and said he never heard of that person. He said he’d never been to the place she lived. The man was irritated. “I’m NOT your father. I can’t help you.” Matt said, “He’s grouchy. He has to be my father.” He figured out what he’d done wrong. He had called the wrong person. He made a phone call. The man was not her father either.
They went back to see what they did wrong. One day he got a new match 649cM that matched on this line. It’s either a cousin or 1C1R or half-nephew? He sent a note asking if he had any Henry or Rutherford ancestors. The guy Matt had just called was his grandfather! He didn’t call back. He decided to go find the grandmother. She said they were married before then and got divorced. She wouldn’t put it past him. She said she’d talk to him—they keep in touch! She convinced him to talk to Matt again.
Then the person who matched 649cM wrote and said, “It’s weird, I got it in my head to do this one day and couldn’t shake it, I really didn’t have some big purpose to do it. Then you happened and I believe 100% that that’s why I was to do it. To help you find what you’ve been searching for and very glad I did.”
He then called the man’s wife. She asked some questions. He said if you ever need to get a guy to do something, let the women take care of it. He said he had a good memory and never met Matt’s mother. He did a paternity test because it’s fast. He told them he might be the uncle. He was not excluded as father. His probability was 99.999999995%. “The alleged father is statistically 7724 times more likely to be the biological father of the child.”
Matt emailed him. He gave him the whole day to look at the email lab report and then called. He said that the test said he was the father. He said he couldn’t talk any more because it was so shocking and gave the phone to his wife and didn’t talk to Matt for a month.
Meanwhile, his other tests came in. His Family Finder is absolutely his father. His Y-DNA test came in and they match 100 markers and 0 genetic distance. He also doesn’t have Y-DNA matches, just like Matt. The only other match is Matt’s grandson.
His father still doesn’t remember his mother. His mother also does not remember his father. Matt figures aliens came down and did it.
They were in the Henry DNA Project.
Matt met his father in October 2016 and then this year in 2017.
Debra did not believe him but he convinced her to take the test. She had a half-sibling match and he identified her father. She finally believes him.
Matt said that one mother gave birth to him for another mother. He finally got closure.
Max introduced Judy Russell, The Legal Genealogist, who presented “After the Courthouse Burns: Rekindling Family History through DNA.” Judy talked about some instances of catastrophic record loss as well as smaller records losses. These resulted in losses of vital records, probate records, and the 1890 census that had partial damage and then they junked the whole thing. To get around this, there are different approaches such as the traditional methodology as outlined in Genealogy Standards. There are alternative records and we can go to alternative jurisdictions. Perhaps when tax records of the county were lost, a copy could be with the state. Think about other places that may hold alternative records and consider alternative people. These are the FAN club—friends, associates, and neighbors, as coined by Elizabeth Shown Mills.
We have genetic genealogical options. This is very much part of the traditional methods in the Genealogy Standards. Standards 12 and 57. Yet, DNA is not listed in the FamilySearch wiki for Burned Counties Research.
In Alabama, there are many county courthouses destroyed by fire. The Cherokee County courthouse burned twice, in 1892 and 1895. Judy’s ancestors lived there. Surviving records are extremely minimal.
Judy shared a research problem in her family that has some evidence but the evidence is not enough to constitute proof for her research problem and there’s no real evidence to prove who the mama is. There’s no direct evidence because it’s a burned county. Using a combination of other resources, among them a divorce and land records, along with a family interview, she had a potential theory.
To answer her questions, she called upon Y-DNA, mitochondrial, and autosomal DNA.
The interview notes said that Judy’s grandmother was a Battles. Y-DNA will not help her. She has an unbroken female line from Margaret. It won’t work for the link to George Battles because his wife has the mtDNA. In order to prove Margaret is a Battles, she needed autosomal DNA. She found someone to match and the match came back perfectly to what she was hoping for. She tested a bunch more and they all matched exactly the way they should for third cousins once removed. With these and many more testers over the years, there’s confidence that Margaret is a Battles and a member of this Battles family.
Who was Margaret’s mother? William had two wives, Kiziah and Ann. Traditional evidence did not provide a compelling case for one or the other, although Ann seemed possibly more likely. To make the compelling case, she needed mitochondrial DNA. From Ann, there was an unbroken female line. She needed one of Ann’s daughter’s lines. In the 1850 census, Ann listed three daughters. There were problems due to records loss. One daughter disappeared from the records. For another, the name was so popular and there many candidates for her. The baby had a really unusual name and she died in Alabama. Her death record lists her parents as William and Ann. It also gave her married name. Judy found them in the 1850 census and they had daughters. However, the courthouse burned and the daughters married so records were not available. Judy sent out letters to everyone who had the youngest daughter in their tree. Finally, a response came back and her husband’s line went back to the daughter with the common name. Judy checked out her tree and it seemed to fit with who she was looking for. When he tested, they were an FMS match with GD 0. DNA helped to pull this family back together!
There’s also an issue of her great-grandmother’s father. The family had some ridiculous stories about a potential father, Jasper Baird. When he died, he left a lot of relatives, as listed in his obituary. There were enough descendants on both sides to use autosomal DNA and determine that Jasper was Eula’s father. Using a21st-century tool in combination with fragments of documentary evidence, using the best practices of our field in Genealogy Standards, it’s possible to rekindle our family history.
I want to leave my results to all of my children? I might set up a trust listing children as the beneficiary of the trust. In her family, the admin of her FF project is the beneficiary of all of her kits that she controls. The admin of her FF project includes herself, a cousin 20 years younger, and they are looking for a cousin 20 years younger than she is. It’s a matter of estate planning.
Did you do a mt test on a descendant of Kiziah? No, but she’s looking for a descendant of Kiziah’s mother to rule it out. They have a theory but not proof.
How did the Battles surnames change? He used Battles in the Revolutionary War. The only time it was Battle singlular was on the one death certificate. There’s a difference in the Y project between the two.
How do you cite DNA proofs? Elizabeth Shown Mills has a Quicksheet available on Amazon plus she has a Quick Lesson that explains how to cite it for free. Read the NGSQ for articles using DNA.
Have you joined the DAR on your Battles line? I am not a joiner.
How much did it cost? First FF tests were $299. She also tested with 23andMe for $399 plus $10 months. Today, 37 tests at $100 each.
Do you recommend admin filing as a limited liability corp.? No. Realistically, think about risk-benefit analysis. Are you realistically facing any liability? If you’re not, then it’s not worth it. If you’re skating close to the line ethically, then something like that might be realistic. If you are following GAP guidelines, then LLC is not necessary.
Did you use a chromosome browser? Of course! It’s FTDNA!
Bennett took a question about Family Finder private projects. He said they’re interested in doing whatever makes sense for admins within privacy restrictions.
Bennett introduced Michael Davila for the Product Update. He is the head of Product Development for FTDNA.
“A user interface is like a joke. If you have to explain it, it’s not that good.”
They are taking a new approach at FTDNA. It’s called user-centered design. It’s based on research about customer needs. They don’t want to build things that have no value or don’t make sense. New users have a lot of trouble. Things like user settings and family tree and anything that’s entry-level is being looked at first.
User Centered Design (UCD) is an iterative methodology that puts the user at the center of all design decisions. Some challenges include accessibility, modern device landscape, and privacy concerns. Some challenges are on tablets. They might make the chromosome browser better for the phone.
In 2017, they had MyOrigins launch, which takes the concept of giving the answer up front and then give a summary of what the data is and then supplemental. They will continue to expand that.
Smooth workflow, no hidden pages
More tips and explanation on why specific information is important, in an effort to encourage people to share more
Better integration of information with the Family Tree
Designate kit manager vs kit owner – they are playing with the concept of having both
Able to upload CSV of surnames
Match and Email Settings
Separate out Match and Email settings, so that users are able to see their full match list but not have to receive notifications about distant matches
Bundling of match notifications
Order tracking and history
New order tracking page, including tracking number visually depicting the transit process and current stage
Clearer order history page, sections for all, open and canceled orders
People will be able to choose a beneficiary up front. If they don’t do any of the parts, they will be sent a reminder email.
The next step for a new user will be to create their family tree. They can upload a GEDCOM file or they can upload a CSV.
The next section is the earliest known ancestors.
Sharing preferences are all on one page now.
New technology under the hood
Easier to create trees for new users
Manage and create multiple trees
Roles and permissions for trees
They did analysis on the performance of the tree and learned that performance of the website is connected to the performance of the tree. They reviewed this and will separate it as needed to make it more functional. The spinning wheel will be a thing of the past.
Trees should have editors, administrators, and contributors.
There are some new options regarding the tree and exporting a GEDCOM.
The new Big Y UI has some improvements. Right now there is the browser that goes full screen and takes you out of your workflow. The new one allows to pull up a panel and search by position and SNP. For some new users to Big Y, there is a bit of context to what they’re seeing. When you’re viewing the results of your matches, instead of having a dialogue box, the people will be on the side and the info will be off toward the right. You won’t lose time clicking back and forth.
Comment: Please make defaults good for projects. A: They want to go with an opt-out method.
Up next is Dr. Caleb F. Davis. He’s speaking about a research project that he’s been asked to do at FTDNA. Caleb worked in precision medicine for ten years. When a family member needed healthcare, his mom didn’t ask for a DNA test. When he came to work at FTDNA, his mom ordered 20 family finder kits!
Caleb was asked to look into the possibility of calling STRs from the Big Y product. Can we call new STRs reliably from NGS and how many? In order to check that what they call with be consistent with Y-111, they want to make sure they have concordance with the truth set and have a high success rate. The matrices should be full, not sparsely populated. They also want to make sure STRs they call today are the same as next week, next month, next year. Sequencing technology is a rapidly evolving field and they want to make sure they are platform independent.
They have 3658 samples with BigY and 111, which gives them what they need to test for high concordance. 1054 candidate polymorphic STRs. They also have one sample they run as an internal control on a few different runs. Lastly, they are in the process of upgrading NGS sequencing technology and want to make sure it won’t change any calls.
For concordance, they had their 3658 samples and there was high concordance with the truth set. They won’t get 111 but they got 72. They compared 263,376 possible genotypes. They used HipSTR and the call rate was 81%. They built an internal caller and it could make calls 85% of the time and with 77%, both were called. This was 99.5% concordant with Y-111. That’s pretty good.
So, then they looked at new ones. They can call Y111 pretty reliably. What about new ones? They looked at about 1000 that were called by both HipSTR and the internal and they ran them.
They looked at the sweet spot and picked about 95% and they were able to reliably call about 393 STRs in 95% of their samples. To make sure they are consistent old to new instruments, they ranked them.
99.5% of samples pass if they attempt to call ~450 new STRs.
The question was how many STRs do you have to attempt to call for the samples to have 289 new STRs. It turns out that you need to attempt to call about 315. There will be about 25 columns that are a little bit sparse, with 289 very reliable.
At this point, as they add more STRs, there are more samples above the threshold very quickly. They may be leaving some useful STRs on the table if they stop at 289.
If they try to call ~450 STRs, it gives about 60 columns that would be a little bit sparse.
What about 600 total? If they go that far, then they’d have to call 600 to get 489 and that starts to get into a less useful reason.
389 for 500 seems to be a reasonable number.
They compared on old and new equipment and were able to get 99.99% concordance across them. There were two samples that were discordant. It was due to not enough coverage.
Why not test all of the Y-DNA? There are 50 million base pairs. A lot of it is repetitive. If you sequence fragments and align them back to the reference genome, you won’t be able to place them reliably.
For new STRs, will it be released to admins? He will work with you directly.
Can you view STR calls on the Big Y browser? Not yet–it’s in the works.
NGS only covers 72 of the 111 STRs. Do you get those missing ones in the larger sets? No. Some of them are simply too long to be spanned. You have to know where it starts and ends and count the repeats in the middle. They use other technologies for those. Those are inaccessible to NGS equipment.
Does it require a fresh new sample to do a Big Y? No, if they have DNA they can run any product. If they don’t have it in storage, then they would need a new sample.
Is this going to replace the Y-111? No. We can’t do all markers in the Y-111 with existing technology. There’s a company called Pac Bio, which is not plagued with short read lengths. Accuracy and repeatability is around 85%. You wouldn’t want to use it to determine wehther you were carrying something important. It’s not ready for primetime. From a technical and reliability standpoint, the prior Sanger technology was really great. NGS produces a huge amount of data and expects bioinformatics people to come up with a solution and sort it all out. The new technology is not as bulletproof as the old technology but the benefit is that you can test huge amounts of data. You could not successfully do a Big Y on a Sanger. You can do work you couldn’t do but there is a technical reliability issue they are very sensitive about. Bennett said if he provides an answer that is not correct, it’s kind of like a sin. Instead of being in a position to ask to reconfirm, they had to take a line and say this is the wheat and this is likely to be the chaff.
After lunch, Maurice Gleeson presented the breakout session “How to Group your Project Members using MPRs: Markers of Potential Relatedness.”
Men with the same surname and a similar genetic signature are likely to have arisen from the same common ancestor, the founder of the surname.
On WorldFamilies.net, the grouping happens automatically. Look at the results tool guide. It compares each member vs Apparent Ancestral Profile, the same as group modal haplotype. Modal value is defined if present in greater than or equal to 75%. The trouble is that extreme outliers will be missed or ungrouped. The great thing is that it automatically updates the project each night. New members are highlighted in green and new results are highlighted in yellow. It handles about 400 people per project, and was only 70 in 2015. Members can also be manually grouped or ungrouped. The caveat is that convergence is not detected. You get a separate page for members’ pedigrees. There are also pretty colors, which help you identify the unique STR pattern for each group within the project.
Terry’s team is Terry, his wife Marilyn, and a part-time programmer. Terry manages hundreds of projects. Please pick a project and volunteer to help.
Another tool that recently came out is Chase Ashley’s Y-DNA Family Grouping App. It uses Y-DNA STR results for a surname project to suggest groups of kits that probably share a common male ancestor and show the genetic distances and relational distances. He uses a slightly different grouping project to WFN. They initially group kits as per FTDNA criteria. No one within a group has a GD higher than 4/37. It’s a very tight group. Then they merge overlapping groups. Most kits will be less than or equal to 4 from the modal haplotype. The max GD between group members can get very high.
Use export to spreadsheet to create a CSV in FTDNA. Open it within Excel and save it as a CSV file.
This app grouped most of them well but the outliers don’t end up included in the group because it thinks they are too far away. It also added Treacy men into the group. These are matches due to convergence, with a common ancestor really about 1,500 years ago.
Maurice talked a bit about convergence. Over the course of the millennia, a marker might mutate in a divergent pattern on various lines and sometimes then a lack of divergence, then back to convergent. They have a common ancestor 10,000 years ago but because they are an exact match today, it looks like they might share an ancestor within hundreds of years when it’s really thousands. There can be back mutations and parallel mutations. Parallel mutations mutate to the same value at a different time in different lines of descent.
Dave Vance did a simulation in May 2017. Parallel mutations greatly outnumber back mutations.
How do we recognize convergence? There’s a higher risk of convergence at lower levels of testing, ie 12 and 25 markers. Also at higher levels of genetic distance, eg 4/27 vs 0/37. When there are lots of matches at 37 (or 67 or 111) markers? There is a higher risk of convergence. Another clue is that there are many different surnames among your Y-DNA37 matches, especially if there’s no particular pattern. Certain haplogroups/subclades have higher risk than other. M222, L226, CTS4466 are more prone.
It’s more of a problem with dissimilar surname matches.
Maurice has talked about this in his YouTube video about Understanding Y-DNA & Grouping People Together.
Downstream SNP testing establishes the terminal SNP or rather, the current most downstream SNP. It helps confirm the correct placement of an individual in a specific cluster and on a particular branch of the Tree of Mankind.
The FTDNA Haplotree is experimental and the most comprehensive. The ISOGG Haplotree is definitive. The YFull Haplotree has dating points for the branches. Alex Williamson’s Big Tree/YTree has surnames and ancestral birth locations at the end of every branch. Maurice would love to see an experimental tree that encompasses all of that. He predicts we will have it in five years.
To do matches’ terminal SNP analysis, open the individual’s matches list down to 25 markers. Sort the list by haplogroup and note the terminal SNPs of his matches and how often they occur. Ignore all the upstream SNPs. Plot SNPs on a haplotree like FTNDA, YTree, etc. DO they all fall on the same or different branches? Look for signs of convergence. Is it occurring relatively upstream or downstream?
Matches’ terminal SNP analysis is surprisingly predictive in many cases. It works best at 25 marker level. Many more matches results in many more SNP tests. Always check higher marker levels first—is there consistency across the different comparisons? It frequently identifies the most likely terminal SNP. It can allow more focused confirmatory SNP testing via SNP pack.
It’s important to tell your project members to join the relevant haplogroup projects. Join mailing lists, facebook groups, and forums. The admins will group you according to STR values. This can help to identify “rare/unique” marker values within group. Never join your members to a group but feel free to email them to tell them that you’d be happy to do that for them and most will write back and ask you to do it.
Looking at a unique STR pattern can tell you whether someone belongs in a group or whether it’s a chance match. Robert Casey has done a lot of work on this and there is a video of this on Genetic Genealogy Ireland’s YouTube Channel.
Elliott Greenspan gave an update.
This year there were 1000 bugs and support tickets ranging from UI to minor data issues. There were over 1000 improvements including speed and reliability issues.
Earlier this year they released MyOrigins 2.0 and nearly doubled the number of underlying populations. They added trace populations and new overview view. Ancient Origins compared autosomal DNA to aDNA. They reviewed archaeology papers for the history of aDNA sites.
Big Y version 2 has been a topic. They have a new matching algorithm. It runs lightning fast in comparison to the old one. The old one took 15 minutes to one hour per sample. As they added more, it got longer and longer. When they added 50 samples at a time with the old one, the new one adds 500 at a time, it wouldn’t work to take 500 hours to upload a new batch. They have done about 50 samples and the new one is down to about 8 minutes. This allows them to make changes to the algorithm and do entire batches at the same time very quickly. In the last one week after the new changes were launched, it sped up quite dramatically.
New user experience focused on the matching screen and how they can show customers the data itself. They are currently showing anyone who is within 30 mutations to the person. The theory is to be able to choose a time in history. Each SNP has a certain amount of years associated with it.
The future of development includes new hardware. How can they take the current hardware they have and enhance it with a new algorithm and for development and horizontal scaling.
New technology includes microservices. They have been completely replacing the underlying database structure. They are looking for consistent load times. There are sometimes load issues on Sunday mornings because they are doing a heavy backup and indexing. This is not good. They are changing the way they do architecture to make things more consistent.
They are looking to have faster updates. They will be able to split work into different segments and have different teams working on different things without colliding with each other. Each little item on the website will be segmented.
The most important thing they’ve invested in is getting high-level talent from industry leading companies. They want to bring in people from oil and gas and tech that have experience with huge infrastructure. This has been their focus and goal and will continue to be.
The 5 SNP tree branching is already rolled out on Big Y. It will make it a little easier, especially in R. It’s difficult to find your branch.
Are you using AWS for scaling? No, they have their entire private cloud inside their facility. Next year there will be a window in there.
When will hg 39 to 38 conversions be completed and when will I be able to get BAM files? 39 to 38 was on the last so probably Monday afternoon it should be completed and then they will start uploading the thousands of samples. At that point, they will be working on releasing the BAM files.
Connie Bormans, the Lab Director, previously gave the Lab Update. They have added a Lab Manager, Brent Manning. He was originally hired as a lab technician. He worked his way up to head of NGS products and earlier this year, Lab Manager.
2017 has been a record-breaking year at FTDNA.
You will soon see that the lab at FTDNA has come full circle. They did this by answering three questions. Did we just go off the charts? How did 40 gallons of hydraulic fluid spill in the parking lot? Are Max and Bennett afraid of heights?
FTDNA was launched in April 2000. Customers ordered Y12, get their kit in the mail, swab their cheek, mail back their sample, and wait for their results online. U of Arizona extracted the data and sent it back to Houston. It was incredibly successful.
They decided the best way to branch out would be to start their own lab in Houston. In 2006, they opened the Genomics Research Center. The first test was the mtDNA launch in 2007. HVR ½ and full sequence were available. They had a DNA sample store, a nice sequencer, and a thermocycler. This is nice equipment but what can you do with only one cycler. He decided to take the plunge anyway and joined in 2006.
They moved to a bigger location at 1445 North Loop West. Hurricane Ike caused an unplanned expansion. The lab was devastated. They lost a lot of equipment. It ended up being a strange adventure. The entire building had to be reconstructed. They had to shut down production or find a way to keep it going. They got generators, filters, plastic sheets, etc. and they set up on the 8th floor for a mini lab. It was big enough for only two people. They worked in there about three months while the building was remodeled. They were successful in keeping production running but they couldn’t produce enough. They got a backlog and had to find a way to do more, better, faster, cheaper. They started looking at modular and scalable pipelines. Processes are broken into sections, or modules. Multiple modules simultaneously and overlap workflows and faster and significantly higher throughput. Standardization provides the benefit of scalability. It proved to be successful.
One of the first big successes was the Family Finder, released in 2010. Part of the success was the fact that is was based on the modular scalable platform. When they needed to do more, they bought more machines and brought in more people. It was dependable with a low error rate. They decided to make the call to bring all the outsourced products to be run in Houston. No longer were they the outsourcee in the lab. This allowed them to ramp up production.
Since they brought it all back home, that’s the point where they had turned 180 degrees.
Part 2 – The Present
The growth has been phenomenal.
In 2006, they did 6 mtDNA tests in a day in a 400 sq ft lab with 15 people. In 2016, they did 384 mtDNA tests per day in 9,000 sq ft with 50 people in just the lab alone.
One of the biggest challenges was Hurricane Harvey. Their staff took a really bad hit. Some lost everything—homes, cars. It was terrible. Everyone came together and helped each other. Everyone in the company helped out with a fundraiser, as did customers and admins.
Another key factor is time. For those in the lab, time was in short supply this year. They were faced with huge projects, sales, and back to back inspections. In September they were inspected by CAP. They passed with flying colors. This past Thursday they were inspected by New York State Department of Health. They are proud to say that they passed without a single issued citation. That is a first in the lab’s history and an impeccable achievement for Connie, the regulatory team, and the lab as a whole. In December they will be inspected by the American Association of Blood Banks. Along with CLIA, this will bring them to four. This allows them to diversify their portfolio.
In addition to genetic genealogy products, they also provide immigration, legal, paternity, whole exome and genome, and copy number testing.
One of the things that has allowed growth is collaboration in a wide range of markets. It is this accomplishment that shows how FTDNA has come full circle.
This path has been difficult. One of the biggest challenges was staffing. Everyone basically ran out of time, were overworked, and people started to burn out. They restructured their hiring program.
INSERT LAB HIRING
They’ve created new positions and put in a much better structure that is allowing success.
- Access to instrumentation
- Shrinking work area
- Throughput limitations
To solve this, they have done a lab expansion. They are taking the entire 8th floor. It’s over 9,000 square feet. They are putting out a fully isolated clean room. This will allow them to do high scale media prep. Construction starts next week.
In addition, the instrumentation. They bought a lot of equipment this year. That one cycler has turned into 19 cyclers. They added 4 iscans, a novaseq, and the list goes on.
Their current extractions were off the charts. The chart went off the graph. They ramped up all of their extraction pipelines. The team broke a record by extracting more than 5,000 samples in one day. They are proud of that sample.
The NovaSeq is a super DNA sequencer. This blows away miseq or hiseq.
In early 2018, they will be providing an upgrade that will allow them to do an update of 4 TB.
In the microarray, they added new HiSeqs. They had two dips—one during Harvey and one when Illumina ran out of chips. They are the third largest consumer of chips in the US.
The LIMS system
- Custom built
- Automated auditing
- Advanced reporting
Manual = Mistakes
Manual tasks have been fully automated
- Reduces fatigue
- Reduces workload
- REDUCES ERRORS!
Some steps that have benefitted are FF pipeline and a lot of tasks on the NGS pipeline.
Part 3 – The Future
The first thing is to elaborate further on automation.
Humans are good at judgment calls, strategy, and theory. Robots don’t tire, are mistake-free, and don’t get bored.
They are trying to take the strengths of the humans and the robots. Let the robots do the boring stuff.
The new thing will be the sample swab and tube. They have updated it and the benefits are staggering. The new sample swab and tube allows them to fully automate all maneuvers. The old one was paper-based and would soon break down in the tube. The problem is that it was difficult to automate removal. They need the robotic tips into the tube. The new product will not break down in the tube. With the orange cap tubes, they can do thousands a day with no one getting a sore arm.
The other new thing is the DNA storage tube. The longterm DNA storage tube is tiny. It has barcodes on the bottom. The ridges on the top make it highly automatable. The biggest benefit of the smaller tube is the storage capacity for different items.
They first had a sample store II and then they switched over to the matrical ministore which held 500,000. The sample store was filling up. The new tube allowed Brooks to make an ultra high-density store. The capacity is beyond anything they imagined. It can punch 25,000 samples per day. The capacity is 2.3 million samples. FTDNA acquired it in 2017 and it’s on its way to being released and launched in the lab. In order to get it into the building, they had to hire a crew to take out the window on the 8th floor and hire a crane to get the nine crates up there.
During delivery, about an hour into lifting they busted a line and the crane crew acted like rockstars. They got a replacement, cleaned up, and gotten a new crane within 60 minutes.
The other benefit of the new sample collection tube is automated sample sorter. Scinomix
- Sample QC’d
- Scanned and sorted
- Placed into rack
- Files created
This will shave three days off the extraction process. This is installed and currently being commissioned. They had a 1200 pound pallet on an 18 wheeler with no loading dock. With 10 guys they dismantled it and brought it to the lab piece by piece. Then they realized the tabletops were bigger than the elevators. They reassembled it perfectly and got congratulations from the crew.
One of the things they want to do is send the boring things to robots and interesting things to humans.
- Move sample in process
- Loading instruments
- QC and approve results
- Reduce change of errors
These are safe systems and they are very team oriented. They are really excited at this frontier.
The lab has come 360 degrees full circle. They are very proud of what they’ve accomplished.
Max and Bennett are not afraid of heights. The open 8 story building was no match for them.
Q&A by Bennett
For Roberta -On a recent blog you mentioned a third-party triangulation tool. What has happened with that?
For Roberta – Please cover advanced matching tools in depth on your blog.
Will Illumina eventually force FTDNA to switch to GSA Array? Yes. They are going to switch to a customized array.
Will FTNDA switch and if so, when? What are the plans to match to current Omni Express? Because they are saying they want FTDNA to do this, it’s going to happen in about 5-6 months. When it comes to how to accurately match back and forth, they’ve got some concessions. They can modify the GSA for more backwards compatibility and more than double the amount of matching content. Max added that in the meantime, LIMS, the program that runs the process, is being implemented.
Will the old red cap sample tubes still be able to be processed? Of course! Even the tethered cap tubes which they still get are also processed. They have to be transferred by hand. About 90% of the samples they are receiving are the new orange top tubes. They will have a legacy transfer issue for years.
Should identical triplets have identical MyOrigins? Connie said this question is likely to be due to the news stories. On FF, it’s based on a microarray. They analyze about 700,000 SNPs. 97-98% of those SNPs will have data. Odds that you will get a result exact every time is highly unlikely. Because they are not analyzing every datapoint, they may not be identical. From a non-lab perspective, Elliott said it runs the sample in replicate many many times. It runs about 20 different replicants and averages the results. Even the exact same sample will have slight variation.
How are you prepared for power outages? Well-prepared. On the lab tour go to the back of the building and you’ll see a huge green box—generator that is hooked up to the lab on the eighth floor. They have already used it more than a couple of times. The lab can continue to run without any issues. They also have backup UPS linked to computer systems. From the moment grid goes down until generator kicks in, there are small backup units to allow power to run continuously.
Comment: FTDNA is broke. FTDNA gets blamed when other vendors change their file formats for raw data. Can you do some PR to change that view? A: Bennett said when chips are changed they don’t immediately import it because of changes.
What is the progress of lab result turnaround time? In 2010 and 2017? Currently huge strides in turnaround time and trying to reduce it to levels lower than currently ordered. Over next few months or quarters, hope to have 2-4 week time due to new benefits, processes, and people.
With all the automation, what happens to the process when a new test is requested on an existing sample? They should already have the sample in their storage. This shaves 3-4 days. With the new store, they can pull and put into process much faster. They expect a decrease in time for those, too.
Can a project have one shared tree for all members? One of the test projects manages to do that. It will be a feature that will be allowed. There might be an option to merge.
Any longer read length expected with NGS equipment? Read length capabilities is not on them. It’s on the providers. Connie said Illumina’s read length is 150-165 bases and they’re not away that they’re looking to change it. PacBio does not have this problem but the accuracy is not comparable to illumine technology. They want to see improvement before they consider it.
Suggestion: On the Y-DNA genetic results display, on matching pages, please add an icon where each item has taken a Big Y. A: Bennett has been asking for that himself.
When I log in and then I go to eat, I log out but there’s junk in the URL and I can’t log out and back in. Shouldn’t it be ready for me to re-login? Elliott said yes, of course, but the way it works is that the junk is due to impersonation. You are logged in as the user and it makes sure you are logged in to that kit.
It’s slower to log in as an admin because it validates that you actually have permission for access to that kit. One of the main reasons is that it validates every time you use an impersonated page. They can look at the junk but that’s why it’s there.
There are three XXX show shirts at the registration desk for someone who has asked for large.
Can administrators download gedcoms from our projects? This got flagged as an ethical issue. If the gedcom is marked as public probably yes but if not public, the answer is probably no. This comes to privacy.
What’s the current cost of a full genome? Max is not going to answer the question but in the same time he can run 800 Big Y or 1 whole genome, he prefers to make 800 people happy. They could do it but for public interest. Connie said she’d buy a new machine. 😊
Does 140 employees include Gene by Gene? They DBA Family Tree DNA. Everyone works for the corporation and are assigned to a division.
When will FTDNA go public? Never if Bennett has anything to do about it. He’d much rather have admins as his oversight than the SEC.
Can you take a family tree offline so that you can send a more updated tree? One discussion is download feature for the gedcom. What will happen is that when you upload, it won’t override. It will have versioning. They are working on processes to link it to the new one and replicate that data.
When will the converted BAM files be available? They will finish V1, upload thousands of samples that are waiting, then begin the process. Probably another 6 8 10 days.
Love the new direct path of SNP tree. Can you add TMRCA to branches? Working on it.
Is this presentation online? Any with permission will be put online end of this week or beginning of following week.
When is next year’s conference? Next year. Probably November. Probably around the second weekend.
Is there going to be a Y-500 test? No, there will be a Big Y and included is everything that Caleb doesn’t think is crap. Bennett doesn’t want to send out data that everyone on the planet is a 4. If it’s not at least a little polymorphic, it can’t give us any hints. Probably in January, they will remove the dependency that you have Y-STRs already on hand to be able to add Big Y. You will be able to buy a singular package that will be Big Y and the additional markers that they are able to tease out.
Will existing Big Y see the STR results in their Y-DNA section or only for new accounts? It’s for everybody! They want to do backward compatibility and not cut out anybody.
On the Big Y STRs, when and how much? When for that as a standalone, January. How much? Don’t know yet.
Max can say how much the Big Y will be when they put it on sale. He will be able to share that shortly.
Will you recalculate all old Big Y? Yes, of course.
Would you consider implementing an API for third-party apps? Maybe, if there’s a benefit for customers. They’ll attempt to write what customers want.
What’s the company’s succession plan? For Max, he’s taking about 10-15 pills every morning and doing a lot of exercising so he can live until 100 years old. Bennett hardly works. He wanders in and answers his emails in the morning after the walk. He’s insistent that Bennett is friends on Runkeeper. The last time Bennett wasn’t working he got in really serious trouble with his wife and he needs a place to go. He plans to keep coming in and puttering around.
When does the holiday sale begin? They wanted the sale to start while we were in the room. They clicked the button while we were in the room. Everyone will receive a copy of the holiday prices.
There are combos. Big Y will be $475 including all the missing markers.
Please feel free to use my affiliate link to make your order. I get a small commission to help pay for my web hosting and maybe even a doctor co-pay for my aching fingers and it does not impact the cost at all.