England GB Groups EIJ

Brian, you're quite right -- only 1-2% of our DNA codes for proteins; the rest used to be called "junk DNA" but now we know that a good bit must have regulatory functions; the latest I've seen suggests that 8-12% "does something" and so maybe 80%+ really is junk! Many molecular biologists have looked hard for functions and explanations for 20 years, and that's where we are. Evidently we large creatures aren't under the same efficiency constraints as a tighty packed virus -- we can happily carry along a lot of extraneous DNA, and that includes all of the SNPs and STRs that we find useful here. We have 21,000 genes and the little water flea Daphnia has 30,000. My favorites are plants: a little Japanese flower has 50 times more DNA than we do. The onion has 12 times more. Lilies have 30 time more. Remarkably this baggage doesn't extract much of a cost, or selection pressure would have pruned it way down long ago. Viruses, and to a lesser extent bacteria, are however under strong constraints, probably just due to their tiny size. I think the argument for lack of function for STRs is just in their statistics. We can account for their properties with pretty simple random models. If they had a function, you'd expect selection pressure (the mechanism of evolution) to exert itself and reduce that variability -- a lot. I'm not familiar with the literature but can guess that biologists have very good statistical tests for selection (which implies a function) vs neutral (which implies no function -- or one that doesn't matter, maybe like eye color). The same holds true for the SNPs we use: if a base pair in the middle of a useful gene gets zapped, then it will usually be strongly selected away (like hemophilia, due to a simple SNP mutation in a clotting enzyme). But the nicely behaved random statistics of SNPs strongly suggest that they're just single-base mutations out in the wide stretches of DNA that doesn't do anything important for our survival.

And according to David Reich, citing others, the era of low-hanging fruit with regard to simple genetic connections to particular traits, conditions, diseases, etc. is about over. Or at least the dream that everything would be so simple is now in decline. Many things are incredibly complex in terms of just how many genetic factors may be implicated and he was frankly pessimistic that humans would ever be able to resolve some of these because of the amount of effort it would take. He labeled the necessary amount of work, a genetic Manhattan Project. The more we learn; the more complex this becomes. And that is nowhere more true than SNP testing.

Oh .... so we are not talking about tree tomatoes (Tamarillos)?

Arthur, that's where AI comes in...for better or worse.

Recycle it my way Denise.

If you are you able to wait that long, and depending on the postage, David, it's a deal!

Thanks for the reply Denise. I might go back to Amazon and read the pieces of chapters they have allowed to see just how well it presents and I understand it. If all good I will go back to fishpond and buy it for their $AU35.00 ...free shipping is their forte I guess. Thanks.

Hey now! You're talking about my family there!

Matthew, how close with FF are we talking about?

5th-Remote cousin. 27cM 15 Longest Block. Not sure of the exact numbers with Michael but it is also 5th-Remote cousin. Michael and I are not Family Finder matches. I haven't had him check to see if the locations are the same yet.

My great great grandmother was Bridget Cassidy (born Higgins) from Galway. This system suggests that you and I have a match at Higgins. I would therefore suggest you start in Galway.

Gavin, were you talking to me (about Higgins/Galway in case our posts get separated by this activity feed)? If so, what system are you referring to when you say "This system suggests that you and I have a match at Higgins."

Robin, well said, and I, for one, do appreciate the ongoing dialog. I certainly wouldn't begin to argue that building a new venture such as this on solid science isn't an admirable objective. It most certainly is...more accurate data and a more accurate interpretation of that data leads to better results. Perhaps then I could confirm if Bob is my long lost brother or uncle or whatever. Unquestionably, some of us will be disappointed going forward. But, until Huxley's "ugly fact" becomes an actual undeniable fact, we must forge ahead with a "beautiful hypothesis" as the best thing going for us. I suppose my bottom-line point here is that much of this should be shared, chewed, and perhaps even digested by the brain trust at FTDNA, et. al. Is that your ultimate goal? If so, keep moving forward but also lets not forget that we have a group of folks here who want to know and feel relatively comfortable about their own "Bob" - within the framework of the results that FTDNA provides.

David, I think this depends upon who is using the data for what, something I would imagine is close to your heart since you have always asked people, "What is your goal in doing this?" Robin, with his usual, admirable, clarity has outlined what he seeks to explore. Perhaps my "bark" conceals the fact that I do care that people understand what this genetic genealogy business is about and receive value for their money. It really isn't hard, but it is hidden in a cloud of terminology that causes people to roll their eyes and sometimes disappear. And is not helped by misleading, ambiguous and incorrect terms that people cannot relate to, or worse, wrongly relate to. No field of science would ever tolerate such verbal imprecision in their practitioners. The way you combined yDNA STR and atDNA data was brilliant and it allowed me to consider similar situations where alternative testing methods might be the most effective way to answer the questions people have. One size does not fit all at the individual level. And around every corner, there could be a surprise waiting. That's the fun of discovery. You never know. But equally, and I know you know this from our past discussions, people have to be prepared for the possibility that this tool will not produce the results they seek.

What I wouldn't give to have mid-morning cream tea and biscuits everyday with this crowd (though, as an American, I am more of a coffee guy)! Thank you for the enlightening feedback. Robin, I did not mean to diminish your efforts in working with STRs (in the event that is how my statements were interpreted), in fact, I am undoubtedly one of your biggest fans. Nor did I mean to imply that there is definitively an "inherent limitation of STR statistics or biology" in general, only that I continue to find that the current models fall short in capturing all of the necessary variables to produce a wholly trustworthy model--which it appears that we are in full agreement. Your analysis on the probability of paternal lines dying out serves as part of the answer to the questions swirling around in my mind (though your explanation on "STR rules" has given me more to think about). When viewing STR data from a modern day perspective, it is hard for the feeble human mind (at least in my case) to conceive how many lines never survived through to modern day. Without these yDNA branches being represented in our samples, it ends up effectively altering STR distributions and misplacing certain mutation groupings (not across the "entire phylogenetic tree" as I previously misstated, but at least within subclade analysis). Regarding trans-111 STRs: I recently read an older article that discussed hand picking certain STRs to serve a particular purpose for research purposes. Could it be that there is a superior way to break up these 500-561 STRs into panels which are specifically designed to allocate all slower mutating STRs to the front and quicker mutating STRs at the end and then layering in each subsequent panel into the analysis (i.e. first applying the panel on slower mutating STRs and then layering in the slower mutating panels one by one) or, alternatively, handpicking a specific panel to address questions that pertain to a particular point in time? Please forgive me if my enthusiasm outpaces my understanding. My attempts to comprehend these things have been somewhat chaotic: one day reading journal articles on population genetics/anthro/forensics/etc., the next reviewing statistics, and the next trying to lay a basic foundation for genetics/molecular biology/chemistry/etc. (the latter being quite daunting for a social science guy).

David, you're right that perhaps these threads go too far into nerdliness -- but I suppose that's up to those who keep them running and "like" the posts. What's funny is your suggestion that there's a brain trust at FTDNA --- I have no personal knowledge but can say that Maurice Gleeson and Iain McDonald would smirk at that suggestion. My sense is that FTDNA's heart is in the right place, but they're just stretched very thin, dancing as fast as they can with an unproven business model (look how 23andMe has moved into medical diagnosis: to monetize its collected data). I'm just happy if FTDNA does well enough to keep this fascinating thing going, and open enough for citizen-scientists to take part. I keep hoping that Bill Gates will get into genealogy and put this all on a Wikipedia-type footing (open, grass roots, supported by philanthropists).