Y-DNA Testing: Where to Do It, Why to Do It, and How Many Markers to Test - Part 3

In Part 1 of this series we talked about where to test, and in Part 2, why to test.  Now let's talk about how to select the optimal number of markers to test.

Part 3 - How many markers should you test?

The simple answer is that you should test as many as you can possibly afford to.

I had a quick look at Family Tree DNA's product page and it looks like they are no longer promoting Y-12 or Y-25. I believe those can still be had for specific testing purposes by joining one of the surname projects, but the current public entry-level test now appears to be a Y-37 at a cost of $169.00. I believe that discounts may be available if you first search for the desired surname project and request a test through the project, but I am not sure. Try that method first, or search online for coupon codes specific to FTDNA or special limited-time offers.

Here's why more markers is better.

I touched earlier on the fact that unique mutations can appear through time within specific branches of a related family cluster. The higher the number of matching markers between you and another tester, the closer in time your common ancestor will likely be found. Higher level tests will help you narrow down a time frame for your mutual connection. In the case of my Henderson example in the previous article in this series, we are upgrading the Delmarva line (currently only tested at 12-marker levels) to compare to our own line which has already tested at higher levels. In this way we may be able to determine whether our common ancestor is likely to have lived before the older group first appears in Virginia in 1661 (in which case the shared connection would likely be in Scotland), or after 1661 (in which case the younger branch likely descends from the older).

Another reason to test at higher levels was mentioned in Part 1 of this series, but not explained. The 12-marker tests will often turn up a number of distracting non-surname matches. You may have shared an ancestor with these matches in pre-historic times, but probably not within historic memory, and in most cases you should not waste time and energy attempting to find the connection. Take my Brooks surname project for instance. At 12 markers, I currently have 151 matches! Imagine if I wasted precious time and resources trying to track all of those folks down just to discover we go back to Og who lived in a cave in Austria! (As if you could ever identify such a distant connection anyway.)

When I upgraded the Brooks test to 25 markers, I was still left with 106 matches, including a large number of testers with non-matching surnames. Too many leads to reasonably pursue. At 37 markers the red herrings and more distant matches begin to fall away, but I am still left with 27 leads. This is definitely a more manageable group, and probably an acceptable starting point for such a common surname, but what if I took things a step further and tested at 67 markers? 19 matches, but at this point, anyone within the same surname is definitely worth investigating further. Even some of the non-surname matches might turn out to be NPE's (non-paternal events) that could actually be genetic Brooks relatives. I typically don't focus my efforts on those, but it doesn't hurt to establish contact. One non-surname match actually suspected he was a Brooks on the male line, and his surname was already familiar to several of us through a maternal lineage. He had an interesting story to tell, but I digress...

By jumping in at 37-markers or higher instead of incrementally upgrading (which can ultimately be more expensive), I can much more quickly and effectively narrow down the matches that are most promising and relevant to my own research, and avoid wasting time on folks who might just muddy the waters to no discernible advantage.

This brings me to the subject of upgrading. If the cost of a 37-marker test is a barrier to entry for you, it is my opinion that it is far better for posterity to get a sample to the lab, even with at the inexpensive 12-marker level (especially for older relatives and most especially for lines that are in danger of daughtering out) than to do no test at all. Tests can always be upgraded in the future as funds become available, but if your potential tester dies, or if a line daughters out as you vacillate over cost, then you have lost a valuable -- possibly an irreplaceable -- opportunity.

In some cases, where there may already be a number of high-level, well-documented testers within a surname, and you simply want to verify your paper trail leading to that family, you can make do with a 12 or 25-marker test. That will be enough to tell you whether you're on the right trail, or barking up the wrong family tree. If it turns out you are a match at those lower levels, then you can always upgrade in order to see how far the match holds up or where mutations begin to appear, and therefore potentially identify exactly which branch you fit into.

So as a general rule of thumb, 111 is better than 67. 67 is better than 37. 37 is better than 25. 25 is better than 12. And for the love of heaven, 12 is better than nothing at all.

Important update 1/1/16: Although I already recommend the use of Family Tree DNA as the test facility of choice, a recent article on DNAExplained.com has raised serious ethical concerns in my mind about the privacy practices of Ancestry and 23andMe. Please read this before choosing to test through either of those two facilities.