A Conversation with Population Geneticist David Poznik

This interview is part of an occasional series of profiles introducing you to some of 23andMe’s scientists. Population geneticist David Poznik came to 23andMe David Poznikafter working in Carlos Bustamante’s lab at Stanford University. Originally from Massachusetts, David earned an undergraduate degree in Biophysics at Brown. An avid traveler and backpacker, David has taken a couple years off to explore Asia and South America. These trips bookended his time as a Senior Research Analyst at Harvard’s Joslin Diabetes Center where, among other projects, he led the analyses for a few early GWAS studies. Upon returning from South America, he went on to earn a master’s degree in Statistics and a PhD in Biomedical Informatics at Stanford. While working in Bustamante’s lab, David became an expert on the Y chromosome. He maintains strong interests in DNA sequencing, ancient DNA, and algorithm development.

Tell us about your education:
I primarily studied math and physics in my first two years at Brown, but I grew increasingly excited about the biology classes I was taking on the side, so I started shifting focus. By the end of college, I was splitting my time between biology and physics, and my research in computational neuroscience was somewhere in between, so I ended up majoring in Biophysics. After college, I conducted the analyses for a number of genetic epidemiology studies at Harvard, but I returned to school in order to gain formal training in statistics and computer science. At Stanford, I earned an MS in Statistics and a PhD in Biomedical Informatics.

What’s your job title and role here at 23andMe?
I am a Computational Biologist and Population Geneticist at 23andMe. Within the Research Team, I work in the Ancestry group. We develop and improve methods to extract insights about customers’ ancestry from their genotypes.

David Poznik, Computational Biologist, Population Geneticist
Originally from: Massachusetts
BA: Biophysics, Brown University
MA: Statistics, Stanford University
PhD: Biomedical Informatics, Stanford University
Fun Fact: “I spent about 25 percent of the decade between college and graduate school backpacking around the world, including a 16-month journey across Asia and a 7-month trip through South America.”

What were you working on before you came to 23andMe?
Before joining 23andMe I was in graduate school, taking as many statistics and computer science classes as I could, while conducting research in population genetics. Most of my graduate research was focused on using Y-chromosome sequences to learn about human history. For example, as a component of the 1000 Genomes Project, our analysis of 1,244 Y-chromosomes revealed multiple historical bursts of extreme expansions in male numbers. Most recently, I developed a pipeline to process and analyze the targeted sequences of several hundred Y chromosomes from diverse African populations.

What are you working on now at 23andMe?
My first project at 23andMe is to develop and test a new algorithm to identify customers’ Y-chromosome haplogroups, with the goal of updating our ancestry reports.

Why are you excited about genetics?
I am most excited about genetics because of the vast amount of historical and anthropological information encoded in our genomes. Beginning with the pioneering work of Luca Cavalli-Sforza in the 1960’s, the field of genetics has made substantial contributions toward our understanding of the past. In some cases these findings have confirmed pictures drawn previously by long-established disciplines such as archaeology and linguistics, but in many cases, and with increasing regularity, genetics has added wholly new elements. Two recent technological developments are driving much of the most recent work: high-throughput sequencing and methods to extract DNA from the remains of people and archaic humans who lived thousands, or even hundreds of thousands, of years ago. Together, these tools have led to a far greater understanding of the history of our species than was thought possible just a decade ago.

Tell us about one of your interesting findings in your previous scientific life.
When I started graduate school, I had the good fortune to have access to some of the first high-throughput sequencing of globally diverse populations. I developed methods to analyze and interpret the Y-chromosome data, and I built a tree relating the sequences. The root of this tree represents the most recent common ancestor (MRCA) of the Y chromosomes in the study, and the branches represent shared paternal ancestry of increasingly smaller subsets of the sample, terminating at the leaves, which represent individual sequences. Thanks in large part to the seminal work of Peter Underhill, much was already known about the topological structure of this tree—the branching patterns that reflect the relationships between the various lineages. But, in contrast to prior work, the full sequence data enabled me to measure the distances between each of the subtrees, or “clades.” These distances are represented by the branch lengths of the tree, with longer branches indicating greater temporal separation. With full sequences, we can infer the time to the MRCA of any pair of Y chromosomes. The MRCA of all extant human Y chromosomes—the root of the tree relating the Y chromosomes of all living men, including carriers of the recently discovered “A00” haplogroup—lived ~275 thousand years ago.

One interesting fact people don’t usually guess about me is:
Something my peers in graduate school never would have suspected is that before returning to school, I actually had a great work-life balance. Since joining 23andMe a few months ago, most of my evenings and weekends have been back at Stanford wrapping up a few papers, but I’m quite excited to start getting out of lab again!

Did you learn anything interesting from your 23andMe results? Did you make any changes as a result of taking the test?
A close friend learned from her 23andMe results that she was heterozygous for Factor V Leiden, and she shared this genetic result with her doctor. In 2010, her doctor found it quite the novelty for a patient to walk into her office, genotypes in hand. She advised my friend that, due to her increased risk of developing venous thromboembolism, she should go off the pill, and my friend complied, reasoning that this was a relatively easy change in light of the potential risk. This event was actually my first window into the power and potential of direct-to-consumer genetics.

What’s one thing the average consumer should know about genetics?
The Y chromosome is remarkable in that it is the longest stretch of non-recombining DNA in the human genome, by orders of magnitude. Because this giant chunk of DNA is transmitted in whole from father to son through the generations, it contains sufficient information to build a detailed phylogenetic tree. At 23andMe, we can tell you where your paternal ancestry fits into this larger context.

  • danodelion

    Swell, first, 23andMe said my paternal haplogroup was I2b1, then, everywhere but 23andMe I was told that I2b1 was renamed I2a2a. Now, 23andMe merely says haplogroup is I-M223, which, some researchers say isn’t an I2 subclade at all, but one belonging to I1, rather, specifically, I1c (read http://www.jogg.info/12/Athey.pdf). I understand that is new and evolving science, but, 23andMe, can you at least please tell us, I-M223 mutants, if we are descended from I1 or I2 ancestors?

  • liz hagen

    Hi David, I started this attached conversation on the paternal updated haplogroup a couple of days ago, however I was thinking you may be a good person to review this as well, since you mentioned the insular nature of the Ashkenazi shtetl Jews. Maybe I should post it on an “Ashkenazi” blog. Wondering if other Ashkenazi also had their % composition recently changed? :

    liz hagen • 2 days ago

    My mother Sophia and I are Ashkenazi Jews with ancestry in Eastern Europe I remember 23andMe reporting us as being almost 100% Ashkenazi- which makes sense. But our top cousins were Catholic -whos grandparents had immigrated from the same Sicilian family – which was so confusing. Now there appears to be a change – my mother appears to be 38% Italian with 48% Ashkenazi. The Italian DNA is recent 23andme says- within a couple of generations. Her top “Sicilian” cousin shares over 4%. WOW. How did this happen? Could this be due to this new update in Paternal Haplogroup assignments? How can I find out how our families intermingled? Hope someone has answers.


    23blog Mod liz hagen • 3 hours ago

    Hi Liz,

    This is confusing to me as well. There is nothing in updates that we’ve made that would have so drastically changed your ancestry composition. The paternal haplogroup assignment changes were not changes to the way we estimate ancestry. They were really changes in how we named and assigned paternal haplgroups. As for the percent shared between your mother and her cousins, that of course is in line with what you’d expect for a cousin, but the description of getting an estimate that so drastically changed does not make sense. I’m forwarding your comment to our Customer Care to help you with your questions.

  • Billie Keaffaber

    Hi David,
    I found my Father Origins I1a1b is it possible with me being female to confirm my relationship to the Brecht Family? I got paper genealogy and documents which put me in family. I finding with the test it picking up here and there on relatives. I picking up my dad mom family Heavy. Lots on my mom side both parents. But unless the hits in 4 generations not picking up my dad dad family is there a reason this could be happening or was I just not given DNA. Most of these test also only go to 6-8 Generations. I am picking up pretty far back on some lines. I like to confirm some the surname links too throughout my tree with DNA I did test that covers all family. Is it Possible to confirm through DNA the different lines through our tree with DNA. I so confused how all this DNA stuff working.

Return to top