Counting Chromosomes
A blog of random musings on genealogy, genetics, science, and history

Why the urgency with Ancestry? Well, we're due for an AncestryDNA shake-up very, very soon.
     —24 June 2019

Update 15 Oct 2019 — Today Ancestry.com officially launched AncestryHealth, entering into the individualized health and medicine arena and competing with their closest rival, 23andMe. AncestryHealth is an offering of two distinct products, one called AncestryHealth Core and the other AncestryHealth Plus, which includes a membership component. For people who have an existing test with Ancestry, the price is a one-time $49 to include AncestryHealth Core, and $98—which includes six months of membership—for AncestryHealth Plus. Both products seem to be completely independent of the genealogy products, and the testing and reporting seem in no way to be linked. Information can be viewed at the Ancestry.com website at www.ancestry.com/cs/health.

Ancestry.com
Ancestry.com is headquartered in Lehi, Utah

Last June I wrote about my opinion that Illumina has probably set a final end-of-life deadline on the OmniExpress chip—the genotyping microarray that the majority of extant results for genealogy have been tested on—and what this might mean for AncestryDNA, the only major company left selling tests based on the OmniExpress chip. Let me quickly state that I have no insider information whatsoever about either Ancestry.com LLC or Illumina, Inc. My opinions are formed only by industry observation and deduction.

I received a few emails about that article noting one line in particular seemed a bit extreme: "Why the urgency with Ancestry? Well, we're due for an AncestryDNA shake-up very, very soon."

Over the past few days, though, we've seen new information which seems to support my previous prediction, and provides a broader backstory to Ancestry's current situation.

Only 20% in Common

SNP Venn Diagram
Diagram showing approximate in-common overlap of tested SNPs between the OmniExpress and GSA Chips

That June post was specifically about what I felt to be Ancestry's imminent switch from the OmniExpress to the GSA (Global Screening Array) chip and how it potentially affects our genealogical research. I don't want to repeat any of that article, but the crux of the issue is that all genotyping microarray chips look only at a select set of genetic markers that amount to about 0.02% of our entire genome, and correspond to about 10% to 15% of the 4 or 5 million SNPs (single nucleotide polymorphisms) that are distinguishing within any one of our particular genomes. From that subset, the OmniExpress and GSA chips test only about 20% of the same markers...or roughly 0.004% of our genome.

GSA Chip
The Illumina Infinium Global Screening Array-24 v2.0 BeadChip

For genealogy, our concern is not knowing which nucleic acids—or more technically, nitrogenous bases—exist at all of our 3.2 billion base pair loci; it's about how one set of test results compares to another. That's how we can evaluate relatedness and build evidence for our family trees. But there are a very limited number of options when it comes to comparing such a small, in-common set of data. These include modifying minimal baseline matching thresholds, the approach GEDmatch has taken, and genomic imputation, i.e., assigning a value to something, by inference only, based upon seeing only pieces of that something. Neither of these options is particularly accurate for genealogy, where we may be trying to provide evidence for a 4th great-grandparent MRCA by triangulating a small chromosomal segment shared by several 5th cousins.

Reasonable speculation is that roughly 29 million people have been DNA tested by genotyping microarray chips, and as many as 20 million of those tests have been done on the OmniExpress chip. AncestryDNA, since it truly took off as a market force, has used nothing else.

Before continuing, I want to be very clear that I, personally, don't feel that AncestryDNA's operations are under any immediate threat. They are not going out of business, and I have just purchased five new test kits for relatives so that they can get the OmniExpress set of results before the chip is retired.

That said, there are several factors in play at Ancestry—some for which it may or may not be reasonable to have expected senior management to have predicted and taken preemptive measures to shift strategy and mitigate risk—and none of those factors are particularly good.

The Ancestry Marketing Model

This will inevitably jump around a bit, but let's start with my opinion of Ancestry's marketing and advertising tactics circa 2016 to present. Perhaps we should say through the beginning of 2019 because, if we follow the ad placements, the "lederhosen or a kilt" ethnicity model was essentially Ancestry's sole push for two years...and they peppered network and cable television with it. We seemingly couldn't watch any channel for longer than 20 minutes without seeing one of those commercials.

Kilt
Traditional kilt, Scottish Gaelic: fèileadh, first recorded in the 16th century

And it worked spectacularly well. So well that others, like 23andMe and Family Tree DNA, felt the need to begin focusing (some) of their own marketing on the ethnic origins component. You've seen those commercials all but vanish in 2019, and we'll circle back to that later.

Ancestry's core business is subscription based: its excellent collection of digitized records and online family trees. And the subscription-based model is a tried and true evergreen revenue generator; we've even seen computer application services like Microsoft's Office 365 and Adobe's Creative Cloud change from buy-it-once to subscribe-and-keep-paying models. I'm certain that the planning assumption at Ancestry was that this "DNA ethnic origins" target market would uncover untapped hundreds of thousands who would become fascinated by genealogy and convert to subscribing members.

The result didn't achieve that. These literally millions of new customers seemed to want that "where did I come from" pie chart and be done with it: 80% Italian, 17% Greek, 3% Eastern European, and they were good to go.

As a corollary, I believe if you ask anyone who uses multiple DNA testing and reporting sites you'll hear that attempting to contact possible cousins via Ancestry yields the worst response ratio of them all. I don't know if it was a failure to educate and cross-sell all those test-takers who were not genealogists, or whether the target market simply never held those cross-sell opportunities.

There is also a high churn rate to Ancestry.com subscribers. Genealogy is an inverse-return proposition. If you have minimal documentation about your family tree, a little research time online can yield big dividends. The more meticulous you learn to be, the deeper and more broad your tree becomes, the more time and effort is required to produce the next new result. So when a casual Ancestry.com subscriber who's been paying and researching for six months or a year gets that new notification that their "All Access" membership is about to renew for the upcoming six months for only $199, some subscribers will simply cancel and start using FamilySearch.org to pick through online source material.

So there's even a sustainability issue built-in to Ancestry's core model. This is a very different subscription foundation than the aforementioned Microsoft Office 365 and Adobe Creative Cloud, for example. I use both of those on a daily basis. They are productivity tools that I can't function without; they're in the budget as essentials. Ancestry.com, on the other hand, is a leisure & entertainment expense. In most households, that budget category is going to be scrutinized every time a new invoice is received.

Pre-2019 Financing and Senior Management

Let's hop next to some history about business structure and senior management at Ancestry.

We do have some net reported income information for Ancestry for fiscal years 2008 through a partial 2013. In thousands of dollars:

  • 2008: US$2,384
  • 2009: US$21,295
  • 2010: US$36,845
  • 2011: US$62,895
  • 2012: US$(1,886)
  • 2013: No fiscal year results, but the Q3 adjusted EBITDA was $54.8 million, compared to $54.1 million in Q3 2012

The negative number in 2012 had much to do with an income tax hit following the Lehi, Utah-based company being acquired by a group led by the European private equity firm Permira Advisers LLP and taking the company private. That restructured ownership arrangement included Permira Advisers as major investors, existing Ancestry management, and Boston-based growth equity firm Spectrum Equity. That deal was completed 28 Dec 2012 for approximately US$1.5 billion. Ancestry's 2012 Q4 total revenue was $131.1 million, an increase of 25.% over the same period in 2011.

In June and July 2013, respectively, Ancestry appointed two new members to its Operating Committee, Janice Chaffin and Brad Garlinghouse. Chaffin spent 21 years at Hewlett-Packard, then nine years at Symantec, retiring as Group President of the Consumer Business Unit. Garlinghouse had been President of Consumer Applications at AOL, CEO of a company called Hightail, and an advisor to Silver Lake Partners LLC (see below). Then CEO of Ancestry, Tim Sullivan, wrote, "We are pleased to welcome Janice and Brad to our Operating Committee."

At year-end 2015, Ancestry announced they had sold "approximately one million DNA kits in 2015; DNA kit sales more than doubled in Q4 year-over-year." A mere augury of the volume to come during the next three years.

But remember their core subscription model and conversion and churn rates? During the period 2012 to 2015 they acquired Fold3, Find A Grave, Archives.com, and spun-off their newspaper-archival component, Newspapers.com, to add to their genealogy-related online properties. So their subscriptions must have been growing rapidly, right?

Note that in the current numbers provided by Ancestry, there is no breakdown of Ancestry.com branded websites versus Fold3, Find A Grave, Archives.com, and Newspapers.com. In the Q3 2013 report, they reported 600,000 paid subscribers from those additional sites; if we keep that same ratio, that would mean Ancestry.com itself now has about 2,400,000 million subscribers. So it seems, in the span of seven years and 15 million DNA tests sold, Ancestry.com has added a net of only about 384,000 paid subscribers. Since they've sold 14 million tests from the start of 2016 until present, that represents a net subscription rate of 136,000 per 14 million, or about one subscription for every 103 DNA tests sold. I'll bet that's not what their marketing plan forecasted.

Majority Equity Acquired by Silver Lake and by Singapore's GIC

News release 1 April 2016: "Silver Lake and GIC Announce Strategic Investments in Ancestry." The first paragraph of the press release reads: "Ancestry.com LLC, the leading provider of online family history data and personal DNA testing, today announced that Silver Lake and GIC, a current investor in Ancestry, have signed a definitive agreement to acquire substantial equity stakes in the company from existing equity holders at an enterprise value of approximately $2.6 billion. Following the transaction, Silver Lake and GIC will hold equal minority ownership positions in Ancestry, while other current investors—the Permira funds, Spectrum Equity and Ancestry management, including President and Chief Executive Officer Tim Sullivan and Chief Financial and Chief Operating Officer Howard Hochhauser—will remain as meaningful equity investors in the company and, along with GIC, will continue to own a majority of the company."

Silver Lake Partners LLC is a U.S. private equity firm, and GIC is Singapore's government-run equity and investment fund.

CEO Movement from September 2017 to April 2018

Ancestry.com CEO Tim Sullivan came to the company in 2005. He had previously, January 2001 to September 2004, served as Chief Operating Officer and then President and CEO of Match.com.

On 12 September 2017, a little over year after the Silver Lake and GIC acquisition, Sullivan stepped down as CEO with no permanent replacement ready. "Howard Hochhauser, the company's Chief Financial Officer since 2009 and Chief Operating Officer since 2012, will assume the role of interim CEO as the company conducts a search for a permanent replacement." Hochhauser remained in that role for seven months, until the following April.

In that press release, Sullivan is quoted: "We've assembled a world-class management team, have an exciting strategic roadmap, and see tremendous potential ahead. I'm also pleased to pass the baton when our company's performance is so strong." Could he have been so prophetically wrong, or did he just see the oportune time to vacate the CEO role?

Margo Georgiadis became CEO at toymaker Mattel in February 2017, and barely a year later, in April 2018 she left Mattel to take the role of CEO at Ancestry.com.

So there's been no real stability at the very top of Ancestry for two years...and these proved to be critical years. Strategic planning and operational contingencies to address and mitigate what is happening now should have been taking place in 2017. They're now boxed into a corner; they're having to be reactive, not proactive.

DNA Sales Growth Slows

DNA Test Sales Growth Rate
Chart by Leah Larkin, from her 22 June 2019 article, "Genealogical Database Growth Slows"

Whatever strategic roadmap Ancestry had in place in 2017 seems to have proven faulty, and just as Georgiadis had the sign on her new office door changed, AncestryDNA sales growth began to falter.

Leah Larkin helps us monitor the pulse of the direct-consumer-testing market; even MIT Technology Review uses her data. As she wrote on 22 June 2019, "The past year has seen a chilling in the genetic genealogy industry: DNA kit sales are down drastically since April 2018." Leah did some rough trend extrapolation and the result was that, from April 2018 to May 2019, AncestryDNA sales saw a (probable) decline in growth of 51%. Similarly, 23andMe seemed to see a 43% decline. If you'd looked at the projections after Q1 2018, you might think we should be over 40 million total tests sold by now. But we likely haven't reached the 30 million mark.

The first, obvious consideration for this that Leah points out is market saturation. Makes sense, certainly in the U.S. where sales have gone on longer than in other countries. And if all I really want is that darned pie chart, you have nothing left to cross-sell me. I personally think 23andMe is better positioned to weather possible saturation than Ancestry.

What Leah calls "the elephant in the room" is no doubt what all you experienced folks out there are thinking. It was April 2018 when information about forensic genealogy and the Golden State Killer appeared, and when we learned that law enforcement had been making use of the matching services at GEDmatch and FTDNA. Misconceptions about "police can see my DNA" ran rampant, and even GEDmatch had people removing their kits from the public database.

Those private equity investors who own the majority of Ancestry could not have been happy about their cash cow suddenly showing 51% lower prospects of milk production.

Now let's add in the probability that Ancestry had been pleading with Illumina, since likely Q4 2017, about the time that Tim Sullivan resigned as CEO, to delay cessation of production of the OmniExpress chip. Those aforementioned investors, operating expenses, and existing debt were vying for every penny that DNA sales were raking in. So now the timing problem that Ancestry should have seen coming. Illumina is absolutely going to stop manufacturing the OmniExpress chip. No other company is still using it, and the money from Ancestry—while the volume might have convinced Illumina to delay the change for a short while—has dried up along with Ancestry's gross sales. It's costing Illumina more money and effort to continue making the chip than the money Ancestry can bring them in chip sales. Illumina is operating at a deficit with the chip, and that can't stand.

Innovation in the Past 18 Months?

Ancestry ThruLines
Ancestry ThruLines Introductory Screen

Within this backdrop we need to consider what has been Ancestry's most significant press release during Georgiadis's administration. Circulated internationally on 28 February 2019 was a release titled "Ancestry® Launches Game-Changing Innovations and Content Additions that make Family History Research Easier and More Efficient." At a time when a true game-changer was actually needed, that isn't what Ancestry produced, despite the hyperbole in the news release. Said Georgiadis, "Today, we are excited to launch several new breakthrough tools and content collections to enable our customers to make more, rich discoveries to unlock their family stories than ever before. It's part of our commitment to continuous innovation as the global leader in family history."

While a legion of people, now with some experience in genetic genealogy, have long been clamoring for Ancestry to, at the very least, display actual chromosomal segment detail (they are the only one of the major testing companies that do not) even if they continue to choose not to provide a rudimentary chromosome browser or mapping tool (again, they are the only ones who do not), what were the game-changing innovations and breakthrough technologies launched? ThruLines™ and the ability to color-code and add a label to DNA matches.

The ability to add a label and a colored dot to DNA matches is hardly game-changing. In fact, it seems rather ludicrous that this would be one of the focuses of a major press release. That leaves ThruLines.

Most every experienced genealogist will consider an Ancestry personal family tree not to be a relevant source in evaluating information for the Genealogical Proof Standard. It isn't that the digitized source information Ancestry offers isn't excellent, and it isn't that there are not mature and very well researched trees on Ancestry. It's because, with the click of a button, inexperienced genealogists can add whole branches from someone else's tree...trees that, frankly more often than not, are poorly researched if not downright false.

Ancestry touts having over 100 million family trees with over 11 billion connections. What we can't know is how many of those contain erroneous data that continues to be copied and propagated over and over again because Ancestry members haven't done their own research, their own due diligence as to whether those 11 billion connections are simply noise, or whether connections you find contain merit to elevate them from noise to the status of information. And Ancestry really can't do that due diligence themselves.

ThruLines™ can show common ancestors customers likely share with their DNA matches and give a clear and simple view of how all matches are connected through that shared ancestor.
     —Ancestry.com

The ThruLines-connected matches being determined, of course, by those 100 million family trees, many of which have big chunks in them that are flat-out fantasy.

Long-time Ancestry.com users looked at ThruLines, tilted their heads, and thought, "Geez; that sounds a lot like Ancestry DNA Circles," which ThruLines replaced. While it may not be a true case of "putting lipstick on a pig" because ThruLines does offer some different functionality, it's awfully close. We still can't compare any DNA segment detail, and the projected ancestral relationships are still driven by millions of trees that we know are unvetted and within which are contained perhaps millions of errors.

Ancestry needed to enter 2019 with a revised, well-considered strategy and innovations that would help secure their market share. What we got instead was ThruLines.

As of a few days ago we now know how Ancestry's marketplace strategy has shifted and where it's going operationally, but before we talk about that let's consider the financial news we saw last Wednesday.

14 Aug 2019: Ancestry Seeks $900 Million Investor Payout

From Bloomberg News Wire, 14 August 2019: "Ancestry.com Owners Aim to Extract $900 Million Payout With Loan." Continuing: "An investor group led by Singapore's sovereign wealth fund GIC and private equity firm Silver Lake Management LLC is looking to pull out more than $900 million from the company through a special dividend mostly funded by new borrowings. They are also seeking approval for another one-time distribution before year-end. The dividend would be one of the largest funded by the issuance of junk debt this year."

This will put Ancestry.com in a highly leveraged position, if in fact the deal goes through as proposed.

Recent declines in DNA kit sales and subscribers have made future prospects more uncertain, making this an inopportune time to raise leverage aggressively.
     —Harold Steiner, Moody's Investors Service

The Bloomberg article describes the terms of the proposal: "Ancestry.com plans to issue a new $1.15 billion seven-year term loan, and use nearly $400 million of cash to fund the recapitalization. This money will be used to make a distribution of around $910 million to shareholders, and repay $600 million of debt. The company's owners are seeking creditor consent to pay an additional one-time dividend of around $150 million at the end of the year, according to Moody's. After that, the company has agreed to use excess cash flow to pay down debt."

Assuming there will be excess cash flow. Ancestry's current debt load, as measured against earnings projections, is a ratio of about 4.2:1. Under the newly proposed seven-year loan, the conservative Moody's and S&P calculations would put the resultant debt load ratio at 7:1, possibly higher.

GSA Chip Coverage
Content summary, Illumina GSA chip; outer ring shows frequency for unique variants present in the 1000 Genomes Project

DNA sales growth is declining, possibly dramatically. Ancestry-provided tools to work with DNA results have been repeatedly and publicly criticized as far inferior to all their competitors. The core subscriber base is growing only minimally. They've seen three CEOs since September 2017. The 14 August 2019 Bloomberg News Network article said, "As earnings kept growing, its owners methodically piled more and more debt on the company to juice their returns." In light of that announced equity/debt proposal, both S&P Global Ratings and Moody's have lowered their outlook on Ancestry.com. And on June 24 last I wrote that I felt Ancestry would be forced to change, very soon, the microarray chip they use for DNA testing.

Make no mistake, that change will cost money, effort, and development cycles. And it isn't a choice that's under Ancestry's control any longer. Illumina, Inc. has wanted to the retire the OmniExpress chip for quite some time.

Living DNA launched 22 September 2016 selling its first test kits in the UK. During the planning process for the new company—as later confirmed by co-founder David Nicholson on the company's Facebook page—Living DNA was advised by Illumina that the OmniExpress chip was being discontinued and that Living DNA needed to use the GSA chip instead, which is what they did, the first DTC company to test with the chip. Similarly, on 5 September 2017 [exactly one week before Ancestry CEO Tim Sullivan resigned], popular genetic genealogy blogger Roberta Estes wrote: "Illumina, the company that provides chips to companies that test autosomal DNA for genetic genealogy has obsoleted their OmniExpress chip previously in use, forcing companies to utilize their new Global Screening Array (GSA) chip when their current chip supply runs out."

GSA Chip Clinical Content
Illumina GSA chip relative summary of clinically-significant markers tested

I fully believe Illumina wanted the OmniExpress completely phased out over a year ago. Up until April 2019, AncestryDNA, Family Tree DNA, and MyHeritage continued to use the chip. Now, Ancestry remains the only holdout. And, especially with Ancestry's sales growth having continued slipping for the past year and their order projections to Illumina no doubt diminishing proportionately, it would surprise me not at all if Illumina's fully-burdened manufacturing cost will not exceed their selling price unless A) Illumina dramatically increases the price per chip to a point where Ancestry cannot remain competitive in the DTC marketplace; or B) Illumina shuts down and repurposes the manufacturing lines currently producing OmniExpress chips, and Ancestry retools its data and matching services to move to the GSA chip.

It would be remiss not to mention that, though seldom disclosed, I believe it is still Quest Diagnostics that does the actual lab work for AncestryDNA, that Ancestry does not—as some believe—have their own laboratory facilities. Among the major for-genealogy companies, FTDNA is unique in that regard.

On 3 August 2016, Ancestry.com entered into "a new multi-year global collaboration" under which "Quest Diagnostics will provide genotyping test services on behalf of Ancestry's AncestryDNA." I have no information that the agreement is not still in place. Quest Diagnostics is a Fortune 500 company, and I imagine it will be seamless and relatively easy for them to change to the GSA chip for Ancestry's processing. The issue for Ancestry will be how to adapt their matching and reporting services to accommodate the different chip.

14 Aug 2019: AncestryDNA Headed into Individualized Medicine

GSA Chip NHGRI Disease Categories
Illumina GSA chip clinically research content based on NHGRI disease categories

Business Insider Prime reported last Wednesday, coincidentally the same day Bloomberg News Network reported on Ancestry's debt-leveraging plans, that "Genetic testing magnate Ancestry CEO Margo Georgiadis said that the startup plans to branch out from genealogy testing and expand into individualized medicine." And, "Ancestry is building out a full health team, with open roles in marketing, engineering, communications, and senior management." This is not the kind of strategic business adjustment you make based on a technology platform that is already past end-of-life. This will absolutely use the GSA chip.

As I said, Ancestry let itself get boxed into a corner. It has to make a change, and the change has to come not of its own proactive choice, or at a time of its choosing. It's going to be far less expensive for them, already being an Illumnia shop using the Illumnia workflow, to switch to the GSA chip. I don't think Thermo Fisher Scientific is even a consideration. But it's still going to require an up-front investment...at a time when sales have slowed significantly and they're trying to leverage debt. Not good. Also not good is that they'll be leaning over into 23andMe's business model, but over two years after that model was implemented.

The Business Insider article said, "Ancestry has sold more direct to consumer (DTC) genetic tests than its competitors—at 14 million—and its massive footprint could give it a competitive edge." Per an Ancestry press release from 21 May 2019, that number is actually 15 million, but I personally don't believe that's very likely to help them competitively.

Remember how Ancestry has marketed its tests, and its essentially flat subscriber base. 23andMe has been focusing on the health/medical market share, and a year ago even inked a US$300 million equity partnership—not simple equity investment—deal with GlaxoSmithKline. Ancestry will have to start chasing for that market from scratch, albeit it does have all of our email addresses.

GSA Chip ClinVar Markers
The Illumina GSA chip includes over 10,000 markers based on the ClinVar database

And there is that continued pesky and annoying 20% SNP overlap in test results between the two microarray chips. Unlike Family Tree DNA's business model, Ancestry stores and archives no DNA samples (I bet they wish they had). All 15 million samples have been tested and the remaining material discarded. They can't go back and offer us all a new test on the same sample. Their only option is to sell us, and everyone else, a brand new test.

Those 15 million existing test results really can't be leveraged to interest a Big Pharma "white knight." The GSA chip was designed from the ground up with health/medical data as a substantial consideration; the OmniExpress chip was not. The GSA chip devotes about 18% of its standard coverage, almost 119,000 alleles, to clinically-relevant markers: protein encoding genes in the exome. Not much help for genealogy, but a lot more relevant to medical and pharmaceutical companies and researchers.

We could see Ancestry...offering a higher-priced test that includes health testing and genetic counseling to guide customers toward treatment options—maybe even from other partner health companies.
     —BusinessInsider.com

While genotype imputation may be acceptable for genealogy and population studies, somebody like GSK isn't going to hand over hundreds of millions for data about genetic markers that are only imputed, or "inferred by association" to be identical. They need specific alleles tested.

 
Ancestry.com is in some trouble. I have no doubt whatsoever that they will survive this; the company is not threatened, and in fact the proposed cash extraction reported by Bloomberg could possibly be an initial step toward taking the company public again.

But it has to reinvent itself...and that reinvention should have been based on proactive, strategic planning that began in 2017. Now the pressure will be ratcheted up a few notches. The next several months will tell us much more about what that reinvention will look like, and how it will affect the genetic genealogy community.
 

Illumina Support Website
Illumina support options for human genotyping microarrays

Added 16:00 GMT 18 August 2019 — On the matter of assuming that Illumina has finally placed the Infinium OmniExpress chip at end-of-life, an astute reader notified me to check the Illumina support website for their microarray products. The URL is https://support.illumina.com/?tab=microarrays.

At right is a screen capture of the options available under "Human Genotyping." We see 10 microarray products listed, and conspicuously now absent is the OmniExpress chip. While there likely are remaining stock that AncestryDNA had previously contracted, this seems supporting evidence that no new manufacturing of the OmniExpress chip will be done (the Infinium Omni5 you see listed is a very different product; it's a four-sample chip that tests over 4.3 million variants).

Thanks to J.H. in Orlando for that timely heads-up.

Edited 12:00 GMT 4 September 2019 — My thanks to all who submitted anonymized AncestryDNA raw data results for analysis in our little, informal project. The submission window has closed, and we will be looking at the 22 kits submitted for various comparisons, including how the iterations of the Ancestry test align since the introduction of v2 in May 2016; how the rsIDs tested compare against both of Illumina's major genotyping microarray offerings, the OmniExpress and GSA chips; and which of the Ancestry-tested SNPs appear in the current ClinVar database.

The analyses will begin soon and should be published by Sunday, September 15. A brief report will be posted on this blog within a day of the report's completion.

My original goal was to have at least 15 kits for comparison, so my thanks again to the volunteers who took their time to prepare, package, and send me 22 sets of results.

—Ed

Edited 23:00 GMT 15 September 2019 — Unanticipated—and time consuming—business commitments and travel as well as what proved to be an unusual way Illumina formatted the latest rsID data for its microarray chips have slowed the analyses. I now expect the report to be completed before mid-October.

As a very brief preview, the 21 sets of raw data we're evaluating (one had to be discarded due to some corrupted information) clearly show that there have been four distinct iterations of the Ancestry v2 chip from its introduction in May 2016 to the most recent revision; the earliest test from the latter we have is December 2018 and the most recent August 2019. The differences are not trivial. For example, for the 22 autosomes the lowest number of SNPs tested among the four iterations was 620,544 and the highest 650,647.

—Ed

See also: