PDA

View Full Version : Better arguments/education needed for NGS



StationMaster
09-28-2015, 07:40 PM
This is a fascinating forum. As a relative newbie much goes over my head, but I still have learned a great deal over the last few months as a passive reader of this forum. I have been very impressed by the group of dedicated citizen scientists and professionals who spend time here. Clearly, a tremendous amount of work goes into furthering our knowledge of genetic genealogy and anthropology. I get the impression that some of you work closely with various testing and analytical companies and that you may be in a position to recommend testing options for newcomers like me. On various forums and email lists I have seen some of these recommendations and I have made some observations from a newbie's perspective. I should also admit that I personally have not yet been convinced to part with $500-600 for next generation sequencing, which is about the cost of a flight from New York to California. Now that is a lot of money, but if I have a good reason to fly to California, I can scrape together the money to do it. If there is a compelling reason to do so, people like me might be able to find money in their budget to do NGS or to upgrade to 111 STR markers.

So, it is with great with respect for the work that many of you do that I offer some suggestions.


Make sure you understand my goals before offering any advice. This is a team sport, but we might not be playing the same game. I might just want to find my birth father and you may be more interested in the Yamnaya culture.
Don't just tell me that I can find my terminal SNP and SNPs that are unique to my family. Tell me how these things can help me. Why should I care what my terminal SNP is? It may be obvious to you, but it isn't to me.
If I have no close matches, help manage my expectations. Will testing more more STRs or SNPs really help me achieve my personal goals? Perhaps I will have to wait until more people test before I see tangible benefits from my own test. That's OK. Just let me know what to expect.
Let me know how my test might help others. This may help me move forward even if the test is not certain to further my personal goals in the near future.
How can my test further scientific goals? Perhaps I am just as interested in the altruistic goal of advancing science as I am in my own genealogy.


Please consider these to be thoughts and questions from a generic newbie. I am not actually asking for advice about my specific situation. I have read on these forums that more people are needed to do NGS, but I think the arguments could be more compelling and provide more disclosure. I fear that some people may expect more from these tests than they reasonably should and that they may then badmouth DNA testing when they are disappointed with the results. Furthermore, I wonder if it is inefficient to tell everyone [and this is an exaggeration] that "xz and z tests will help, but NGS and 111 markers is always best". Maybe my testing budget could be best spent on a donation to someone else's test rather than my own. If I have $500 to spend, would it be better for me to do Big Y or to test three distant paternal cousins at 37 markers? Perhaps my money could be best used as a contribution towards the NGS of another tester whose results are likely to provide more critical results than my own? I am not sure that I am actually that altruistic, but I am sure some people are.

I like the Lewis and Clark analogy, but this only explains how to best find the "holy land" when you have no idea where it is. This analogy assumes that everyone already wants to go to the holy land and they know why it would be great to get there. We know why Lewis and Clark wanted to get to the Pacific. I suggest that you take a step back and help newcomers like me understand why we should want to find our terminal SNP in the first place. Why is this the "holy land"? Many thanks.

Táltos
09-28-2015, 08:48 PM
Welcome to the forum StationMaster! You have put forth a lot of good points about testing NGS. I am sure you have read various threads on the forum about the same. Ultimately it is up to you what your goals are. Initially you might start out just wanting to know about genealogical family, but then you may decide that you are interested in deeper ancestry.

Money is always a factor. Great analogies about the plane fare. People have all different things to take into consideration about this. If you have the money for the advanced testing up front, and don't suffer from getting buyer's remorse, and your significant other won't give you grief because they don't understand your hobby...I say go for the more advanced testing. That's just my humble opinion. :)

miiser
09-30-2015, 09:09 PM
The value of NGS needs to be assessed for each individual tester. In some cases, there is the potential for genealogically relevant information to be learned from a NGS test. In other cases, the potential does not even exist. To generalize that NGS is the best option for every tester is simply not true, and to recommend NGS testing to every genealogist is bad advice.

As a starter, I assume that (because we are talking about genetic genealogy, and not anthropology) the goal of the tester is to discover genealogically relevant information, specifically, the identity of their ancestors, where and when their ancestors lived, communities, tribes, or clans their ancestors were associated with, the origin of the surname progenitor, and their relationship to others sharing the same surname.

If the interest is in anthropology, then NGS of the typical modern tester is unlikely to reveal anything of scholarly interest (beyond the current state of knowledge) for the overwhelming majority of testers, who are mostly Americans from the British Isles.

In contrast, this implies that the tester is not particularly interested in knowing their terminal SNP, except so far as that information is relevant to the genealogical information listed above. This also implies that it does make sense to "chase someone else's SNPs" rather than to discover one's own SNPs. The goal is to learn genealogically relevant information regarding the relationship to other testers, not to discover trivial information about the location of mutations occurring within the DNA of one's own private lineage.

Based on this foundational principle, I suggest that the correct approach to assessing the value of NGS testing for a person is to start by asking, "What testing has already been done? What is already known about this person's placement within the tree? What is known of the relationship to their nearest matches?"

Once this has been established, the follow up question should be, "What are the possible and probable outcomes of an NGS test? What new information can possibly be learned from an NGS test? What are the specific questions needing to be answered regarding the tester's genealogy that have the potential to be revealed by an NGS test? Is an NGS test likely to reveal any new information regarding the tester's ancestry beyond what is already known from previous testing? Is a lower cost test, such as a single SNP or SNP panel able to answer these specific questions equally well?"

In some cases, the answer will be that NGS can reveal genealogically relevant information for the tester. In many cases, the answer will be that previous testing has already answered all the questions that can possibly be answered regarding the tester's position within the tree and their relationship to other testers. In these cases, NGS offers no genealogical value to the tester.

For most testers, the main limitations on what can possibly be learned from genetic testing are:

1) The limited database size of people participating in genetic genealogy. By this I mean not only those who have done NGS testing, but those who have done any sort of testing at all. NGS and SNP discovery refines branch definition. But, for many testers, the primary limit on the information that can possibly be gained from testing is the limited database size. Additional branch definition is often not needed to understand the nature of the relationship to the other testers within the database. In some cases, even if all the people who have done any kind of Y-DNA testing were to take an NGS test, nothing new would be learned of a person's genealogy. For some, STR and targeted SNP testing have already answered all the questions that can possibly be answered from the limited sample size.

2) The limited knowledge and limited documentation of the lineages within the database. Getting a match to a tester who knows nothing of their own history reveals nothing of your own history.

Branch definition and refined phylogeny, which is what NGS Y-DNA testing has the potential to reveal, is a limiting factor only for some subset of testers.


Some will argue that, with NGS, you are testing for posterity, for the information that will hypothetically be learned in the future. Along these lines, they will say that NGS testing is a team sport. But the "team sport" argument is also sometimes used as a carefully glossed way of arguing that it's morally acceptable to talk other people into spending money that won't help them discover their own genealogy, because it serves our/my agenda.

An NGS test might in some cases be reasonably recommended with the expectation that nothing of genealogical relevance can possibly be learned from today's database, but for the much larger hypothetical database of the future, the SNP discovery will provide useful information. But in these cases it is important that the tester be aware that the test is recommended only for its hypothetical future value.

And I would counter that the cost of NGS testing is likely to be much lower in the future. In cases where the currently small database size limits the potential for anything new to be learned from NGS, the best strategy is to put NGS testing off to the future when the cost will be lower. When the time finally arrives and there are genealogically relevant matches within the database, and NGS testing is needed to refine the phylogeny, a person may be able to do an NGS test for $100. It is foolish to purchase a $500+ NGS test today for a hypothetical future value, when the NGS test may be purchased in the future at a much lower cost. The exception to this is when the only available DNA sample of a lineage is in a person who may soon be deceased.


Some will also argue that NGS testing has a lower total cost than sequential piecewise STR and SNP testing. This may be true in some cases. But in many cases, as little as a 37 marker test is sufficient to learn all there is to be learned from the database. In some cases, an upgrade to 67 or 111 markers may be sufficient. In some cases, targeted individual SNP or SNP panel testing may be sufficient. For many testers, incremental testing will answer the tester's genealogical questions at a lower total cost than NGS testing.

For some testers who are not budget constrained or are in a hurry, it might make sense to go directly to a conclusive NGS test, and reduce the risk and time of incremental testing. But FTDNA does not offer NGS to new customers. The test is only available to people who have already purchased an STR test. For someone who already has an STR test, the tester should evaluate the outcome of the STR test before considering an NGS test. The NGS test should be pursued only if it has the potential to answer specific questions that the STR data cannot. For someone who has not yet done an STR test, and is pursuing NGS for the purpose of saving money by eliminating incremental testing, they should choose a product that doesn't require you to first spend money on an STR test as a prerequisite. FGC is a better option in this case.

NGS testing is a good investment for some testers. But for other testers on the opposite end of the spectrum, a 37 marker Y-DNA STR test is sufficient to reveal all the genealogically relevant information that can possibly be obtained from Y-DNA testing. Many testers do not have a good enough understanding of genetic genealogy to evaluate what may possibly be gained from their own NGS. They rely on an unbiased project administrator to evaluate the situation and give appropriate advice for the best value to the tester. NGS testing should be recommended only when it is likely to answer a specific question about a person's genealogy that lower cost testing is not able to answer.

Táltos
09-30-2015, 10:45 PM
Miiser all good points. I have found with "chasing other's SNPs", that in hindsight it would have been better to test my brother at Full Genomes so I could personally have the most definitive answer. True not enough have tested to be in databases yet. All I know is my brother's closest 67 marker match and him do not share the same terminal SNP. That man does share that SNP with others though, so it's not like it isn't genealogically relevant.

miiser
09-30-2015, 11:03 PM
Miiser all good points. I have found with "chasing other's SNPs", that in hindsight it would have been better to test my brother at Full Genomes so I could personally have the most definitive answer. True not enough have tested to be in databases yet. All I know is my brother's closest 67 marker match and him do not share the same terminal SNP. That man does share that SNP with others though, so it's not like it isn't genealogically relevant.

Yes, there are such cases where the STR markers do not clearly point to the correct group, for sure. In my experience, though, they are not typical, and not common enough to recommend everyone do an NGS test from the get go. The average cost of finding the "terminal" (genealogically relevant) SNP in my projects is something like $250. The cost tends to be higher for lineages from less heavily sampled areas of the world, where there aren't so many matches to be had from the STR test.

In your example, with the products that are available today (assuming you were starting from scratch and didn't know your terminal SNP), would a 37 marker test followed by an $88 SNP panel have done the job?

The discussion is also complicated by the fact that most of the individuals encountered by the people giving the recommendations - people in this forum, or in FTDNA projects, or in Yahoo discussion groups - are people who have already invested in an STR test. We are, unfortunately, not the first encounter most of these people have with DNA testing. If we could have caught them at the FTDNA front marketing page before they spent a dime, we might have stopped them and told them to go straight to FGC for NGS. But FTDNA's marketing strategy seems to be to lure people along from one test to the next, one or two hundred dollars at a time. And we are not usually in a position to give advice to people until after they have already purchased an STR test. Once the STR test has already been purchased, the cost saving advantage of NGS is no longer applicable.

Táltos
10-01-2015, 12:48 AM
Yes, there are such cases where the STR markers do not clearly point to the correct group, for sure. In my experience, though, they are not typical, and not common enough to recommend everyone do an NGS test from the get go. The average cost of finding the "terminal" (genealogically relevant) SNP in my projects is something like $250. The cost tends to be higher for lineages from less heavily sampled areas of the world, where there aren't so many matches to be had from the STR test.

In your example, with the products that are available today (assuming you were starting from scratch and didn't know your terminal SNP), would a 37 marker test followed by an $88 SNP panel have done the job?

The discussion is also complicated by the fact that most of the individuals encountered by the people giving the recommendations - people in this forum, or in FTDNA projects, or in Yahoo discussion groups - are people who have already invested in an STR test. We are, unfortunately, not the first encounter most of these people have with DNA testing. If we could have caught them at the FTDNA front marketing page before they spent a dime, we might have stopped them and told them to go straight to FGC for NGS. But FTDNA's marketing strategy seems to be to lure people along from one test to the next, one or two hundred dollars at a time. And we are not usually in a position to give advice to people until after they have already purchased an STR test. Once the STR test has already been purchased, the cost saving advantage of NGS is no longer applicable.

A 37 marker, plus the the $88 SNP panel would have given me the same answer I have today. At a cheaper cost though. :) My thing is I really am not sure how does Big Y compare to Full Genomes for Q1b. I honestly cannot answer with certainty if there are any Q1b men who have taken both Full Genomes and Big Y to compare what is actually being missed by Big Y. Or perhaps I should say, I don't know if there are any men that could be from the Q-YP1003 branch that have taken Full Genomes. They have only done Big Y. I know some men who are from the Q-YP2197 branch have taken Full Genomes, some did Big Y. So not sure if they are able to compare what's missing or not among the Q-YP2197 branch between the two companies.

My brother's 67 marker match went on to be Q-YP1010, as did other men. My brother was negative for that. He is positive for Q-YP1003, which the 67 marker match has as well. The other man in the project who is only Q-YP1003 only matches my brother at the 12 marker level. There is also another man, who is only Q-YP1003. He has been testing at YSEQ. Off the top of my head I'm not sure if he did STRs there too. His first test was at 23andme, and it told him he was a Q1b. BTW, he does not know his family of origin on his direct paternal line. But that's okay in my book. I do not mind helping out someone who is looking for their roots. He may turn out to be a closer relative to me than I had first realized.

A lot of people did STR testing at FTDNA, because that was the only thing available once upon a time...

TigerMW
11-04-2015, 09:53 PM
...
So, it is with great with respect for the work that many of you do that I offer some suggestions.


Make sure you understand my goals before offering any advice. This is a team sport, but we might not be playing the same game. I might just want to find my birth father and you may be more interested in the Yamnaya culture.
Don't just tell me that I can find my terminal SNP and SNPs that are unique to my family. Tell me how these things can help me. Why should I care what my terminal SNP is? It may be obvious to you, but it isn't to me.
If I have no close matches, help manage my expectations. Will testing more more STRs or SNPs really help me achieve my personal goals? Perhaps I will have to wait until more people test before I see tangible benefits from my own test. That's OK. Just let me know what to expect.
Let me know how my test might help others. This may help me move forward even if the test is not certain to further my personal goals in the near future.
How can my test further scientific goals? Perhaps I am just as interested in the altruistic goal of advancing science as I am in my own genealogy.

This is a great list, Stationmaster.


... Maybe my testing budget could be best spent on a donation to someone else's test rather than my own.
This often is the case. My experience is this doesn't go over that easily as it just doesn't feel good in many cases for the donor. Pooling resources in a project contribution system to spread the costs and then using a preset criteria or lottery to pick an NGS test winner seems to work okay.


..I like the Lewis and Clark analogy, but this only explains how to best find the "holy land" when you have no idea where it is. This analogy assumes that everyone already wants to go to the holy land and they know why it would be great to get there. We know why Lewis and Clark wanted to get to the Pacific. I suggest that you take a step back and help newcomers like me understand why we should want to find our terminal SNP in the first place. Why is this the "holy land"? Many thanks.

I couldn't find where I posted this but have a an old copy. I generally preface this with the concept you brought up - that it's important to know what one's goals are because the recommendations are altered based on the goals and abilities so "it depends." I agree entirely.

The preface of the Lewis and Cark analogy is related to genetic genealogy. It sounds like a discussion of what genetic genealogy is, is important. I don't think the goal is to find one's terminal haplogroup, per se, but to connect one's known genealogical family into the tree of mankind and understand closely related familes, origins, etc. and in some cases, older pedigrees. I wouldn't confuse genetic genealogy with a holy land but what do you think it is? or how should it be described? We have to have some generalizations and terms that are generally applicable. I can see that some of these terms are not intuitive for sure and may mean different things to different people.

------ Lewis and Clark Expedition analogy for NGS --------

I've been looking for an analogy for Big Y and Next Generation Sequencing (NGS) in general. I propose the analogy of Lewis and Clark's expedition to explore the U.S. Louisiana Purchase lands with an aviation twist.

Lewis and Clark's primary objectives included "explore and map the newly acquired territory, find a practical route across the Western half of the continent." They left St. Louis in 1804 and arrived at the Pacific Ocean late in 1805.

In this analogy, we can think of the Pacific Coast as our genetic genealogy homeland, the place or status where our genealogically known family connects specifically to the Y DNA tree of mankind. The Pacific Coast is not the same for all of us, as each of our families has a distinctive location. Lewis and Clark founded what would be Fort Clastrop on the edge of Astoria, Oregon. From Astoria Column, a tower, you can see the Pacific Ocean, the Cascade Mountains and the Columbia River.

Single SNP testing is like flying a two seater from St. Louis and hoping to land in Astoria without knowing where Astoria is. The plane is low priced and reliable but has bad gas mileage. More importantly, Astoria may not even have coordinates on the map yet or a landing strip. This kind of approach is most applicable when someone who is highly probable to be on the edge of your genealogically known family has already done a Big Y test and has built a very tall tower or lighthouse to go with a new landing strip. That tower in Astoria could be thought as a super version of Astoria Column and it is built with 111 Y STRs.

Fixed SNP panel/pack testing is like flying the two seater from St. Louis hopscotching across the country, landing at a handful of small airports and getting out and taking a good set of photos at each location and then deciding the next location to fly to. Fixed SNP packs/panels are a good, low entry price way to go, but suffer the same problem any fixed SNP test suffers. What if your Astoria hasn't been discovered? Perhaps, even your State of Oregon has not been discovered. You also might have troubles if your eyesight or navigation system isn't so good. For good navigation you'd want to have at least 67 Y STRs although 111 is a better long term investment.

Discovery testing, like Big Y, is like having a super high speed, fuel efficient jet traversing back and forth on multiple paths high across the sky on mostly clear days taking special photos of the countryside between St. Louis and the Pacific Coast. It is scanning over 11 million locations. If your Astoria turns out to be San Diego, Long Beach or Tacoma, that's okay. Big Y is accomplishing what Lewis and Clark were doing, mapping the route for settlers to follow in the form of lower entry price tests. Unfortunately, your family of genealogical record might not even be on the maps for the mass migration of settlers to come, that is without Big Y discovery testing. It's just a fact of the Y chromosome just as it is of the geography. The settlers won't go to a place when they don't know where it is or even know it exists.

Only a member of your genealogical family can discover your Astoria and erect the Astoria Column of 111 STRs for the settlers.

I'm asking you to start thinking about Big Y if you haven't already. There are at least several thousand results now in. It works. Pooling of resources at the project/family/surname levels can help.
https://www.familytreedna.com/learn/...testing/big-y/
https://en.wikipedia.org/wiki/Lewis_...ark_Expedition
https://en.wikipedia.org/wiki/Astoria_Column

-------------------------------------