Genetic Genealogy Testing

If not familiar with the terms STR, SNP, DNA, Chromosomes, Autosomes, X, Y and Mitochondria; then skim this page now until you read the next section that covers these general terms of genetics in more detail.

There are four main types of genetic genealogy tests; shown in general order of importance to you and the project. There are two more we discuss below but, for various reasons, are pulled out separately.
  1. Autosomal with X (Somal) SNP,
  2. Y (Somal) STR,
  3. Y (Somal) SNP,
  4. Mitochondrial SNP,
  5. Next Generation Sequencing (NGS),
  6. Autosomal STR, is briefly introduced for completeness but is not part of genetic genealogy testing.
It is very important to understand the differences in these tests as the testing companies do not clarify their offering and make claims contrary to what you might expect for the type of test performed. Each type of test is explained in more detail here. See the DNA glossary or simply click on the terms for further clarification of each. Or read the next section DNA Primer first for a more in-depth explanation of genetics before trying to understand the Genetic Genealogy Testing section here.

Quick Summary

The tests and coverage offered by the major USA-based testing companies are below (with the project MINIMUM recommended solution in turquoise). International offerings are constantly changing and we encourage you to consult with the admins about your case:
DNAType23andMeAncestryDNANGG Geno2.0 NextGen 5 FTDNA (non-USA Only)FamilyTreeDNA 1FullGenomesUse
& X
SNPYes ($99) GOODYes ($79)Yes ($150) GOODFamilyFinder ($79) GOOD2Whole Genome7 (Best) ($1,425)Finding any relative with an MRCA in the last 200 years. X included in RAW data file for all but only visible / analyzed in 23andMe
YSTRnononoyDNA: 12, 25, 37, 67 and 111 markers ($59 to $170 to $360):4300 markers; yes or yElite 2.16 ($795)MALE-ONLY; Near term (<1,000 years), patrilineal relatives (male, surname line)
YSNPincluded, GOODincluded, OKincluded; GREATBigY ($575)3 + yFull ($50) (BEST), or
SNP Pack ($120) (GOOD), or
per SNP ($40 each)
yes (in both)MALE-ONLY; Ancient patrilineal Haplogroup (>5,000 years); Verify same STR family (BigY is NGS test for nearer term relatives and lines)
MitochondrialSNPincluded, GOODnoincluded, OKmtDNA tests ($70 to $200; including full sequencing)yes (in both)Ancient matrilineal Haplogroup (>5,000 years); Verify matrilineal line. Minimal genealogical value.

1 FamilyTreeDNA offers a separate product, for a separate cost, for each testing area. A-la-carte. 
2 xDNA SNP results are only seen in the FamilyTreeDNA FamilyFinder Raw Data file; not on the website. Ditto for AncestryDNA.
3 BigY is not available unless you have already ordered a yDNA STR product. BigY ($575) with yFull ($50) can be used to determine most yDNA STR markers as well; but the STR results will NOT appear in your FTDNA account or the FTDNA group study chart.  BigY is not reliable for all STR markers. As of April 2015, FTDNA now removes Autosomes, X and mtDNA “minimal, essentially useless” results from the BigY BAM file before providing it. BigY, along with the yFull $50 paid additional processing to extract all the STR’s, will place you in the yFull phylogenetic tree.
4 A yDNA STR 12 marker test can often be used as an enticement (free, introductory gift kit) to show if the tester is “in the right ballpark” as the group they are comparing too. A 12 marker result can also be a minimal cost way to take the all-inclusive BigY test. As of 2016, it seems the only way to order a y12 or y25 kit is by phoning in OR one at a time from a Project Join Page when not logged in.
5 National Geographic Genographic (NGG) Geno2.0 NextGen, which was updated in 2016, now has a full Autosomal with X test included that can be transferred into FamilyTreeDNA as well as GEDMatch and others. This test has the best yDNA SNP testing outside of the BigY test from FamilyTreeDNA. You will always want to transfer your NGG yDNA SNP results into FamilyTreeDNA for the fee. NGG’s Autosomal results can be imported indirectly back to FamilyTreeDNA by modifying the RAW data file to appear like an AncestryDNA file and then doing the normal AncestryDNA file transfer. There is a small fee at FamilyTreeDNA to make imported FamilyFinder discovered matches available to you. A similar mechanism can be used to load the NGG Autosomal results into GEDMatch. Note: as of 2017, in the USA, the NGG test is supplied by Helix and markedly different in the amount of autosomal data in the test. Helix results cannot be imported into either FamilyTreeDNA yDNA SNP database nor the GEDMatch Autosomal match database.
6 yElite 2.1 is an NGS test that surpasses doing FTDNA: y12 STR, BigY, and mtDNA testing.
7 Whole Genome is an NGS test on ALL your DNA: the Autosomal , X & Y, and mitochondria. The test returns over 1.5 million SNP values; more than 3x from the nearest autosomal test competitor. Due to the depth of testing, it returns phased results as well — very important to the accurate matching as well as helping determine which line the match is from when the parents DNA is not available to test.

NGG Geno2.0 NextGen is a very close runner-up to 23andMe for the recommended solution (but outside the USA only now). For males AND those wanting to stay with “one company”, it can be a better test due to their extensive yDNA SNP coverage and the ability to directly import the yDNA SNP results back into FamilyTreeDNA. NGG was originally developed and administered by FamilyTreeDNA. But you cannot use DNA samples already stored with FamilyTreeDNA (or vice-a-versa). Note that the 23andMe Autosomal match database is much, much larger than either FamilyTreeDNA or GEDMatch. NGG does not offer any match databases or DNA test analysis tools. NGG in the USA is now from Helix and does not meet these criteria.

As of Spring 2016, we have reverted back to recommending 23andMe as the entry-level test. It is now cheaper (again) and has so many other features that outweigh the others. In Fall 2015, due to the price doubling at 23andMe, we changed the recommendation to a hybrid of AncestryDNA and FamilyTreeDNA as the lowest cost, entry point. But AncestryDNA’s (a) requirement of a subscription to do anything useful with the results and (b) lack of chromosome analysis tools is a problem for most. And FamilyTreeDNA’s charge to import Autosomal results also gave pause. 23andMe still has good analysis tools, includes the mtDNA test, and, importantly, has good yDNA SNP coverage for males. You are always encouraged to use GEDmatch for analysis of Autosomal results. If you want to buy as many sample kits for as cheap as possible and then later order different tests against the collected samples; then use FamilyTreeDNA’s a-la-carte solution by ordering y12 kits through the project. You can modify the y12 kit to something else before sending the sample in (for females). See our test comparison table below for more details.

Autosomal with X (Somal) SNP

Autosomal (atDNA) with X (Somal) SNP tests can be used by ANYONE to find ANY OTHER relative if you both share ANY ancestor within about 6 generations. That is, if you have a common grandparent born since 1800 or so. This test is called FamilyFinder at FTDNA and simply the product at 23andMeAncestryDNA, and NGG. While a surname project like ours does not support finding autosomal matches directly, each testing company website provides support for matching your results with those tested within that company. AncestryDNA requires their search service (subscription fee) to contact matches or see matches trees. NGG does not support a match list but your results can be transferred to FamilyTreeDNA to obtain the full services there.  GEDMatch is a third party site, created by the RogersDNA surname project like ours, that allows you to upload your autosomal with X results from the testing company and compare them to others who have similarly uploaded; possibly finding more matches from testers at different testing companies. AncestryDNA and NGG do not provide detailed, numeric match strength information nor a chromosome by chromosome segment match analysis. 23andMe, FamilyTreeDNA and GEDMatch all provide this. (It should be noted that NGG does not even provide a match table / capability with other testers at all.)

Positive match results are likely if you have a common ancestor within 6 generations of another tester. More distant relatives may still indicate matched but the “norm” for fifth or sixth cousins is generally below the noise level and testing limitations, for European people and cultures.  The quick drop-off of matching ancestors is due to two issues: (1) getting only 1/2 your DNA from each parent each generation, and (2) crossovers (aka Recombination) during Meiosis replication that breaks apart longer matching segments of DNA in a chromosome. Generally, there are no more long, matching segments of DNA still remaining between 5th cousins. As recombination is statistical, shorter segments tend to stick around longer. So sometimes, you can find a “sticky” matching short segment 10 generations out. Not often, but sometimes. Third cousins or closer are near guaranteed to find matches with each other, when tested. 4th to 6th may find useful matches. 6th to 10th cousins may find a match, but rarely. 10th and beyond cousins will rarely, if ever, match Autosomal and X DNA.

Results are often reported as a list of matches with others who tested at the company.  How they form the list and the order it can be given is a mystery sometimes.  But generally the strongest matches are listed first.  The strength of the match (how much of your Autosomal DNA matches) indicates how closely you are likely related.  None of the companies guarantee a “match” on your list will respond to an inquiry to explore your relationship further.  GEDMatch and FamilyTreeDNA are the only ones to provide an actual email address of your match. Unlike reported elsewhere, our experience is ALL the testing companies have about the same customer matching response rate to inquiries. Some companies provide more information and analysis tools than others; sometimes only available after a match agrees to share further.  This can be the numerical percentage you match, the amount of your DNA you match, and/or the specific segments on specific chromosomes you match with someone else.  Matching of DNA segments of is determined by looking at all the SNP markers checked in the test, finding contiguous runs of identical values, and then determining how much of the chromosome length that sequence of matching SNP markers represents.  RAW results are reported as individual SNP marker values — either changed or not; sometimes as deleted or inserted.  The more SNP’s tested on every chromosome, the more accurate the match results.  The length of a matching segment, the number of matching segments, and the number of chromosomes with matching segments all contribute to helping determine how closely related you are.  The longer the segments matched and the more segments on more chromosomes you match, the more related you are.  There are some populations or conditions in your past where this may make the results more indeterminable. Any culture or time where there is a great pedigree collapse due to marriage of close cousins in the recent past or frequently over many generations fit this criteria. Parsi’s, Ashkani Jews, and Mennonites are examples of such populations or cultures.

Phased results, if possible to determine, help to remove false-positive matching segments (that arise due to the testing process) and to determine whether a paternal or maternal side is contributing the matching segment.  Reportedly AncestryDNA phases their results and so greatly reduce false positive matching.  But our experience shows they have the highest rate of false positive matching. (Reported weak matches on AncestryDNA will show no match at all when both are transferred to GEDMatch.) You need a parent’s results to compare with a child to phase the child’s results. GEDMatch and Immanuel provide phasing tools.

Sometimes you only see a summary percentage of the match with someone else instead of a total length of all matching segments.  Length is often expressed as centiMorgans (or cM). As a very rough rule of thumb, divide the cM total match length by 72 to get the percentage match; and vice-a-versa.  So 72 cM of total matching segment length is a 1% match.  Note that other than the RAW results of actual SNP marker values, the result can only be reported as an amount matched or in common with someone else. Some include the X chromosome matching in the percentage or total segment length match result while others do not.

Autosomal SNP tests are commonly used to report on the ethnicity makeup or ad-mixture. That is, the (historical) percentage makeup of your ancestry from different regions of the world.  This is done by identifying certain, more unique marker values or sequences of values associated with populations or regions before more recent migration and mixing. As the Autosomal SNP test looks across 44 chromosomes, there is more variety and mix of your recent background represented than solely with a Y (Somal) SNPMitochondrial SNP or X (Somal) SNP test. So the accuracy is much higher and represents your real mix.  The accuracy of this percentage makeup is getting better every year but is still the subject of much debate. Percentage ethnicity makeup is most accurate when using as much, if not all, of the DNA as possible.

Most companies providing the Autosomal SNP tests simply include the X (Somal) SNP test results as well. FTDNA only reports the X (Somal) SNP results in their RAW data file and does not support analysis of X (Somal) SNP matches with others. 23andMe and AncestryDNA also include the Y (Somal) SNP results when a male tests and use that to determine your ancient Haplogroup. They do not include Y (Somal) SNP matches in their match analysis.  23andMe even includes the Mitochondrial SNP test and thus casts the widest net of coverage over all your known DNA. They are the only single, inclusive test covering all your SNP markers in all the DNA and thus reporting on all your ethnicity, patrilineal haplogroup and matrilineal haplogroup in addition to recent relatives you match.

A Parent-Child relationship is always guaranteed to share 50% of their chromosomes. The actual percentage can appear slightly off 50% depending on whether they include the X and/or Y chromosome result in the comparison process. Siblings share 50% of their DNA, on average, but it can vary (theoretically) from 0 to 100%.  Testing companies throw the half- versus full-IBD matching under the rug and simply report, at most, 50% sharing. In fact they never report the nearer 50% match as they only do half-IBD matching which generally only reports around 40% matching for siblings. The standard deviation off the roughly 50% matching mean among siblings appears to be around 10-15%.  Because of this variance in sharing, siblings reported ethnicity make-up can appear quite different. See our own Consanguinity page or the ISOGG page on Autosomal Statistics for more information about the expected sharing of Autosomal DNA with relatives.

X (Somal) testing results are often just lumped in with Autosomal ones, if mentioned or used at all. But we remind you, as you delve deeper into refining your knowledge and results, to look for how to separate out and analyze the X (Somal) results independently. There is some additional information that can be gleaned from this special chromosome that has some characteristics of the Autosomal chromosomes and some of the Y (Somal) chromosome. The cross-over (recombining) regions at the ends of the X (Somal) and (XandY|Y)) (Somal) chromosomes are often reported as part of the X (Somal) result. AncestryDNA reports this recombining region as a 25th chromosome in their RAW data file.

Because Autosomal chromosomes dilute by 1/2 with each generation, and because statistical variation of how much is actually transferred varies ever more widely with each generation of separation, we strongly urge all members to get as many older relatives tested as possible. And the more on the more branches, especially branches further back in your ancestry, the better. This will allow you to triangulate matches with others to narrow down the likely branch where you are matching someone. Testing multiple siblings is a refinement but with different benefits than trying to get distant cousins from as many branches tested as possible.

  • The transfer of Autosomal results from 23andMe (v3 kits from 2012 and 2013 only!) or AncestryDNA to FTDNA throws out all X, Y and mtDNA results.  The transfer cost of $39 is worthwhile to get more matches from others who tested on FTDNA.
  • Our experience shows 23andMe’s Haplogroup and ethnicity analysis to be most believable.  But which company is better at reporting your ethnicity changes as each conducts and/or incorporates the latest research.
  • Since early 2016, AncestryDNA has the largest database of testers; more than double all the others combined. 23andMe has the second largest database of Autosomal testers, and then distantly followed by FamilyTreeDNA (with their FamilyFinder test), NGG, MyHeritage, and (for completeness) GEDMatch.
  • There are usually hundreds of thousands of SNP markers tested across the Autosomal and (Somal|X)) chromosomes. How many SNP’s tested varies by vendor and has been captured by us in a table.
  • It has been reported that FamilyTreeDNA removes the medical-indication SNP’s from their results file before providing the RAW data file to the user.

Y (Somal) STR

Y (Somal) (or yDNA for short) testing can only be done by males and then to only test if they are related to other tested males when both share a common male ancestor along a male-only ancestor line. Females need to find a male sibling, cousin or the like to take the test.  The STR test covers Microsatellite (aka Short Tandem Repeat or STR) markers that, by design and selection, change more often than SNPs. This test works for upwards of 1,000 years (30 generations) and is not limited to the 6 or less generations of atDNA SNP testing.  The Y (Somal) chromosome stays intact, for the most part, and does not have cross-over during meiosis or multiple parents contributing similar segments. The yDNA is passed from father to son mostly intact; following a patrilineal line. For most of western Europe, the patrilineal line will follow the surname inheritance and hence the proliferation of Surname DNA projects like this one among the Western European population. This is the main test for participation in this project.  The test is for looking to match patrilineal descendants from common male-line ancestors. This test indicates when a biological link exists, or not, through the last 500 years or so.  Most male lines have not been characterized nor analyzed so finding a match and learning from it is often a long, detailed research project.

Y (Somal) STR testing is by far the largest component of FTDNA’s user base.  The 12 marker test is useful to see if you are in the ballpark with a surname project; at a lowest initial cost.  A 37 marker test is really a minimum to verify you are in a surname family. More than 37 markers help determine branches in a family.  Sometimes more testing is needed as marker values are very close between different surname lines in the population. Markers can change value and even change back, surname lines determined by markers can become overlapped and confused after a thousand years or more when patrilineal lines diverge then re-converge in their STR marker values. Hence, companion yDNA SNP testing is a requirement as well.

You can start with a 12 marker test and upgrade later at FTDNA. Historically, there is no additional cost at FTDNA for testing incrementally; just additional time to reach your full results. Note that you are testing the Y chromosome only; which is passed from father to son virtually unchanged with no crossover (in the tested region) during meiosis replication.  Hence, the chromosome is very stable through the generations. With that said, STR markers are chosen that are less stable than traditional SNP ones and thus change more frequently. Often one marker change every hundred years out of every 30 markers. Hence the more markers you get tested, the better the comparison and matching that can occur.

Results are reported as a count for each STR marker tested. You should match all markers identically with siblings to 2nd cousins and be different by one in a one or two marker values for 3rd to 6th cousins; when using 37 or more markers to compare.  Roughly, one marker value off by one for every 125 years in 30 markers compared till the MRCA. Some markers change more frequently and markers can change back.  Sometimes a marker changes length (count) by two at a time. So the estimation is not absolute. In fact, a tester may have a unique marker value that even their father does not share.  STR results can be used to predict your Haplogroup.  But SNP markers are the defining way to determine a Haplogroup. Confusingly, an STR test result defines your Haplotype; not your Haplogroup. Similar terms with very different meanings.

You can work with the project admins to map your yDNA STR results into the subgroups already identified or let them help take everything you have and figure out if there are any new avenues of genealogical research for you going forward. One of the project admins was stumped twice.  Another admin helped them get past the name change around 1900 that was a complete surprise.  And then DNA testing helped find the second name change at the 6th generation in 1800 by linking him to a 12th generation ancestor. Working forward from that ancestor drastically narrowed the search space to find the stumped 1800 link.  So check out our pages and see if our collected genealogical research and DNA results might help you.

Order a yDNA STR test in the Hoar / Hoare / Hore Surname DNA Project and receive a discount. Already tested at FTDNASimply join the Project.  Were you yDNA STR tested at another company?  Transfer your results to FTDNA to join this project or contact the admins to provide your data outside the FTDNA portion of the project.  The yDNA STR test is somewhat unique to FTDNA but is also offered by GeneBase.com, Oxford Associates and YSEQ.  See ISOGGs’ yDNA STR Testing page for more info. Beside the matching inside the FTDNA company site / database, FTDNA also supports a free-site named Ysearch.org for members and non-members to register, upload and compare their results. This is similar in idea to the GEDMatch site for Autosomal test matching.

Y (Somal) and Mitochondrial SNP

Y (Somal) and Mitochondrial SNP testing can be used to find your early human ancestry along a particular line — male-only or female only — back tens of thousands of years.  This is often categorized as a Population Study or Haplogroup / Phylo Tree understanding.  As mentioned earlier, Y (Somal) is abbreviated yDNA. Mitochondrial DNA is abbreviated mtDNA. Like for yDNA STRyDNA SNP only tracks your male or paternal line; that is patrilineal.  Similarly and conversely, mtDNA SNP tracks your female-only or maternal-line (i.e. matrilineal) as the Mitochondrial DNA only survives from the egg.  The Y chromosome is inherited by males from their father.  The mtDNA is inherited by everyone from their mother.

SNP markers that are tested, by design, will more rarely change and thus are commonly shared among many people. As such they are not as useful for nearer-term genealogy and surname work. Those with common yDNA or mtDNA SNP results are not necessarily related in any near term line. This is more so for mtDNA because it is, by far, the shortest DNA strand and consists of only 16K base pairs. There are very few SNP’s to test. Do not be oversold here. yDNA and mtDNA SNP are not used to find relatives in a genealogical time frame. Only early ancestors thousands of years ago and that only those that share your patrilineal or matrilineal line. A very small part of your overall ancestor background. With that said, their is use and value in the project for these tests. Especially now that NGS tests area available for yDNA. And especially for yDNA which has many more SNPs to test.

The yDNA SNP markers are used to help solidify the STR results. If your yDNA SNP values are not matching, then even if you have similar yDNA STR values, you are not related in the last few thousand years; if at all. yDNA SNP testing helps with the yDNA STR results in the surname project by confirming your haplogroup grouping and that the matching results are valid. Those with very similar yDNA STR results must have near identical Haplogroups as determined by their yDNA SNP results.  STR results can sometimes predict, but not determine, the Haplogroup.  Matching yDNA STR results but different yDNA SNP results mean your STR match simply happened to collapse to the same values over time with others; but that you are not related in the last thousands of years or more.

mtDNA strands are very short (only about 16.000 base nucleotides) with very few SNPs.  Thus, it is very common to have the same mtDNA test results but no relation to any other matches in the genealogical time frame. Two people with the exact same markers likely share a common matrilineal ancestor thousands to tens of thousands of years ago. Two people with a common matrilineal grandmother should match their mtDNA SNP marker values. But two people with the same values are not necessarily related in any near-term time. So if you have two testers who have a common grandmother along their matrilineal line, then you can compare their mtDNA SNP marker values to see if they are identical; as expected. This is the only real benefit of mtDNA testing to genealogists. mtDNA, like for yDNA SNP testers, have Haplogroups defined for their results. So you can determine ancient family membership for your matrilineal line.

Because  SNPs are often stable for thousands of years, those with common SNP test results may not be related in any genealogically determined time frame. Because  SNPs do randomly change though, deep SNP studies in surname projects can sometimes identify changes that occur in the near-term at a particular branch point. Infrequent changes do not mean they cannot change at all in the near-term.  The once every 1,000 or 5,000 year SNP change could have occurred last week.  Identifying these private SNP changes can sometimes help determine family lines in a surname group just as STR values can. This has been happening with regularity as the NGS tests become available and more widely used. This only applies the the yDNA as there are over 50 million base pairs with well over 50 thousand SNPs to possibly test and compare. mtDNA is just too short with to few SNPs to find any real value in the hear-term (except for the before mentioned narrow case).

Instead of individual yDNA SNP tests at FTDNA, project admins recommend testing with 23andMe or NGG initially. With the latter, you can import your yDNA SNP and Autosomal results into FTDNA as FTDNA developed and perform the NGG project testnote 1.  FTDNA only supports importing atDNA and not yDNAxDNA nor mtDNA SNP results from all except from NGG.  AncestryDNA provides some yDNA SNP results in their test but no mtDNA. NGG reports far fewer mtDNA results than 23andMe but often enough to determine your maternal Haplogroup.
note 1: As of 2017, in the USA, the NGG test is provided by Helix and not FTDNA. It is unclear if this import capability of yDNA SNP results still exists.

A phylogenetic tree for mtDNA is maintained at PhyloTree but no tools exist to map your results into the latest tree available.  FTDNA offers MitoSearch as a free, companion site for anyone to upload mtDNA results from any company and compare it to others. But, as mentioned, the utility of this is just not there for genealogy.

Next Generation Sequencing (NGS)

NGS is available and extremely useful to the genetic genealogy community. While a catch-all marketing term that covers many different techniques, it basically describes testing of the DNA that starts to more closely resemble full sequencing. The cheap, genetic genealogy tests described to date use test array chips, or assays. A very over-simplified idea is they return ink blotches and you are comparing an ink blotch result to reference samples. Once you make the closest comparison, you make a call on a large number of SNP values. OK, way over simplified, but you get the idea. NGS is more fine grain and tending to pick up individual base-pair values and thus finding more SNP changes that have been overlooked till now. Instead of thousands to tens of thousands of SNP’s in a chromosome, you get ten’s to hundreds of thousands of results per chromosome. Enough coverage that many of the STR values are reliably extracted from the results returned. Full Sequencing does exist but is still prohibitively expensive and only used in medical research environments.

FamilyTreeDNA’s BigY test is an example of NGS. It appears, by name and claim, the NGG Geno2.0 Next Generation test may be an NGS test also (to be confirmed). Full Genomes has been the organization offering the NGS service to the genetic genealogy community for the longest time. To make an investment in BigY useful, you really need to use the yFull site (another $50) to do a full interpretation and extraction of results. But realize yFull results cannot be reimported back into your FamilyTreeDNA account.

It should be noted that these tests, to date, are not as much about Genealogy as they are about pushing the frontiers of genetic genealogy. But they are starting to bridge a gap between Ancient (>5,000 years) Haplogroup studies and nearer term Genetic Genealogy results to find distant cousins. Especially when applied to the yDNA chromosome, NGS can be used to do a more extensive SNP and STR analysis that begins to link surname lines together (generally, when the MRCA is >500 years ago). This is bleeding edge work that is filling gaps of Autosomal testing at the fringe of genealogical studies (300 to 600 years).

Autosomal STR

Autosomal (or atDNA) STR is more commonly termed DNA Profiling or DNA Fingerprinting. The technique is looking for a few STR markers across the 22 Autosomal chromosomes. This technique was the first invented and is predominant in criminal forensics and paternity testing for legal and court claims. Databases kept by governments of criminals, or evidence collected and tested in crimes, falls into this category. Although yDNA STR exists for genealogical purposes and usually requires a large number of markers, the criminal / legal test uses only Autosomal chromosome STR markers and chooses more highly variable STRs then would be useful to the genealogical community. Hence they can use a much smaller sample or tested marker count for the diversity and matching they achieve. No genealogical test company offers atDNA STR tests and hence they do not overlap with criminal or legal testing technology. This should relieve some concerns when recruiting relatives as their results will not end up in a governmental or criminal database for comparison.

Notes on Testing

  • 23andMe gets the most for your money on SNP testing.  They provide All DNA SNP tests in their single product for $99: Autosomal, X, Y and mtDNA. NGG is next best at ((SNP) coverage with $150 (plus a possible $39 transfer fee to FamilyTreeDNA) getting you All; but has weak mtDNA coverage although a much stronger Y coverage. NGG offers no match database capability. AncestryDNA is next best at SNP coverage with $79 getting you Autosomal and X but minimal Y and no mtDNA. (AncestryDNA also provide the least analysis and numerical reporting on their website and require a subscription to their service before you can contact any match or see their genealogical tree.)  In FTDNA, their separate a-la-carte testing requires you to pay $745 or more to get all the SNP’s on all the DNA.  Hence, we recommend FTDNA for yDNA STR testing only and the other companies for everything else.
  • The transfer of Autosomal results into FTDNA throw out the somal (X and Y) and mtDNA SNP results that might be in the RAW data file (except with NGG where they only capture the Y and throw out the Autosomal and X; go figure?). FTDNA only accept the latest AncestryDNA tests (2014 or later) and v3 23andMe results (roughly 2011 to Oct 2013). But you can modify 23andMe and NGG result files to appear like AncestryDNA ones and get them in, if motivated. 23andMe drastically changed the SNP mix with their v4 as did AncestryDNA in 2017 both reducing the overlap with FTDNA and others of the SNPs captured. Thus reducing the number of SNPs available for comparison across companies. This mostly affects GEDMatch that accepts input from all the services.
  • 23andMe: has only recently started saving samples and it is not yet clear what it means as they only offer a single test. FTDNA has been saving the samples sent in since their beginning around 2000.  One of our admins ordered an additional, newly offered FTDNA test eleven years after the initial test / sample.  The relative had long since died but the admin was able to get the benefit of the new testing technology. So something to consider when deciding vendors.
  • FTDNA’s reported database size is mostly based on yDNA STR only testers.  Their SNP tester DB, for an apples to apples comparison and for FamilyFinder matching success rate, is believed to be 1/10 to 1/20th the size of AncestryDNA or 23andMe’s match database. 23andMe testers, historically, bought the product for health indicators and not necessarily genealogy. So although you may find matches there, actual responses to genealogical inquiries are sometimes sporadic. Or the people’s knowledge of their family history may be sparse as they are not into genealogy. For best results, do autosomal testing directly at all the companies. As of 2016, it appears AncestryDNA has passed 23andMe (in terms of number of Autosomal testers) and well on their way to double 23andMe’s test base size and reach 4 million testers by early 2017.
  • Our experience shows 23andMe’s ethnic background analysis to be the most believable of all the autosomal test companies. But this is rapidly changing with all the companies and they poor resources into this, the biggest reason, that the general public tests.
  • FTDNA: charges separately for each and every yDNA SNP marker (result) and each type of SNP test; you have to order them individually. BigY is a recent addition to help cap the cost of yDNA SNP testing when you are not sure which SNP’s to test. But you have to have already ordered a yDNA STR test before they make BigY available to order. So although they offer the most extensive DNA testing, in general, they are very costly for the coverage you get. And the benefit to your personal genealogy is limited. They are more for beyond genealogical time frame analysis and use. The recent yDNA SNP Packs introduced are a way to help make their yDNA SNP testing more inline with the other solutions.
  • X chromosome results are lumped with Autosomal by the testing companies mentioned here; if mentioned at all on their website. Only 23andMe makes them explicit and analyzable on their website.
  • SNP tests are, by design, looking for unvarying single (or small sequence) base-pair changes; often within genes themselves. SNP variations in a gene can be an indicator for health issues.  STR tests on the other hand, by design, look for more frequently changing repeat sequences in the inter-gene material that makes up a majority of your DNA.  That is, outside the genes themselves.  STR values generally change within a few thousand years whereas SNP values change many tens to hundreds of thousands of years apart.  This is by definition and design.  So STR tests generally help prove you are related within a genealogical time frame.  SNP tests can prove you are not related, if very different results, but say nothing when results are the same (except you historically, thousands of years ago, shared an ancestor). This time frame of estimation to a common ancestor is shrinking with NGS testing.
  • Some helpful charts on STR values are on Leo Little’s page. Also, a useful backgrounder on DNA testing is on Kerchner’s page. See our FAQ on Genetic Genealogy Terms in the Forums for some additional information. According to an article by Felix, the BigY test from FTDNA includes all 480+ known STR regions as extractable data. And beside Felix’s tools, yFull offers to do this extraction for you. So if you plan to eventually do BigY, limit your STR testing as it will be covered there.  See ISOGG’s Autosomal DNA Statistics page for interesting descriptions of SNP values and their expected results. Tim Janzen did a great job with the (Autosomal) SNP testing comparison chart.
  • Recent efforts (summer 2015) were posted by a few. Blaine Bettinger doing an informal survey on Autosomal SNP results that gives people’s experience on SNP segment lengths in Autosomal testing.  (Unfortunately, we have since found major falacies on the data collection process, documented in our article.) There is a nice post by Jim Bartlett on why it is not just a blanket coverage (variation exists) when comparing SNP’s of known cousins separated by 5 or more generations).

Comparison of Major USA Testing Companies SNP coverage

We have been documenting an accurate table and comparison of the SNP testing capabilities of each company (and test version as it changes over time) as we did not find it existed anywhere. Here is what we have collected so far:
Number of SNPs Tested AncestryDNA (2015-)10 23andMe (v2 -2010) 23andMe (v3 2011-13) 23andMe (v4 2014-)6 NGG Geno2.0  NGG Geno2.0+ NextGen (non-USA only) FTDNA8 FamilyFinder Affymetrix before 20122 FTDNA8 FamilyFinder Illumina since20132 FTDNA8 BigY NGS3 FTDNA8 mtDNA4 Helix (2016-)11
ALL 701,478   ? 960,613  602,352  142,131   ?  ? 708,193  60K+9  ? TBD
Auto+X 700,593  566.905 7 956,388  596,869  130,109  716,005   ? 708,193  TBD
Y 885   ? 1,766  2,329  11,978   13,533  25K+ 3 TBD
Y in ISOGG5 233 (42)  ? (?) 333 (71) 425 (69) 1,837 (151) 1,265 (141) 1,162 (57) 3 TBD
mtDNA  ? 2,459  3,154  441 411 1,143 4 TBD
Cost $79  $100  $100  $99  $200  $150  $499 2 $79 2 $575 3 $225  $99

1 Need to investigate. This may be reporting just the values that are changed from the standard reference model; instead of values for each SNP tested like the others
2 FamilyFinder test (autosomal-only on web; X included in RAW data file)
3 BigY test covers all known yDNA SNP markers and those not yet unidentified (25,000 to 38,000 SNPs). Before April 2015, FTDNA used to include 18,765 atDNA+X and 72 mtDNA SNP’s (Total: 71,355) in the BAM file. They were misleading and useless and now simply stripped before delivery. The price dropped $100 then also. yFull does analysis to extract many of the 111 STR marker values. Some report that 450 STR marker values can be extracted. FTDNA does not do any analysis to extract STR values from this test.
4 mtDNA Plus covering full sequencing of the HVR1 and HVR2 coding regions shown; they offer full sequencing to get the complete 16K base pairs of the whole mtDNA strand,
5 Number of Markers tested that are in ISOGG (in parenthesis, the anecdotal number of ISOGG SNP marker’s showing a Positive Change for a few testers that were checked). 2014 Tree and SNP list used for comparison as common basis; before major expansion.
6 23andMe’s newer test has reduced the overall number of Autosomal SNP markers but increased the yDNA and mtDNA test marker count. In Oct 2015 they doubled the price to $200 and started offering some health indicators through their website again. In Fall 2016, they offer a new Ancestry-only $99 test.
7 Values from Felix Chandrakumar / Immanuel and his work on comparing Autosomal tests (note: 2014 and earlier posts at InternetArchive fc.id.au, 2015 is captured from the new site Internet Archive fi.id.au.)
8 FamilyTreeDNA has separate products for each test type. BigY, in the RAW BAM file, has included some Autosomal, X and mtDNA results in the pre-April-2015 test but no longer.
9 BigY is a quasi-full sequencing approach (as opposed to an array chip used in other tests from FTDNA and most other companies). So counting SNP’s extracted is a bit of a misnomer as more information is really in the RAW sequencing file.
10 AncestryDNA requires a subscription to the search service before you can do much useful with the results on their website. You can download the RAW data file without a subscription and thus upload into GEDMatch though. You cannot contact any match or see their genealogical tree without a subscription. As of Winter 2017, it appears AncestryDNA may be changing to a subscription service to even see the matches, let alone contact them.
11Helix is a new startup in the field formed by Illumina (the company building the testing equipment), National Geographic Genographic (NGG), and others. Starting in 2017, it appears the NGGs Geno2.0 +/NextGen test in the USA is the Helix test; which provides a very different data set, no access to RAW data, and no longer importable to FTDNA. Outside the USA, it is still using the FTDNA provided NGS test.