What Happens After I Spit? A Saliva Sample’s Saga

The way our Personal Genome Service™ works is pretty straightforward, at least from a customer’s point of view. We send you a saliva collection kit, which is at its heart a plastic tube. You spit in the tube and send it to our laboratory, which extracts DNA from your saliva, analyzes it and deposits the resulting genetic data in your account.

Though the process usually goes off without a hitch, each sample has to clear a number of quality control hurdles in order to be processed. If it doesn’t, a customer’s data may take a little longer than expected.

Sample problems can almost always be corrected one way or another. It is only in very rare cases that a person’s saliva contains so little DNA that repeated efforts to analyze it fail (In those situations, we do provide a refund!).

For those of you who really want to know all the messy details, below is a detailed account of what can go wrong with a sample, and what we do if it does:

Step One: The Visual Exam

Samples arrive at our lab every day. The first thing technicians do with each new sample is scan the barcode on it, because that’s the only identifying information the laboratory has. For privacy reasons, the customer’s name and other account information never passes beyond the 23andMe firewall.

As soon as the samples have been scanned, they undergo their first test — a visual inspection to make sure they contain fluid up to the “fill line” on the side of the tube. Occasionally samples do not pass, either because they leaked in transit or the preservative solution (buffer) isn’t released into the tube when the customer snaps the oblong cap onto the collection funnel. Either way, if a sample tube doesn’t contain enough liquid we immediately contact the customer and ship a free replacement kit so he or she can spit again.

Step Two: DNA Extraction

If a sample passes visual inspection, it’s time for DNA extraction. The process our laboratory uses to isolate and purify the DNA in your saliva is simply an automated version of the same one you can do at home that allows technicians to process batches of 96 samples. Technicians remove a portion of the saliva from each tube, then put the rest of the sample aside so they can try again in case their first attempt fails. If a second attempt to extract DNA from a particular sample fails as well, we send a second kit to the customer who provided it (free of charge) so that person can provide a fresh (and hopefully DNA-rich) one.

A sample may lack enough DNA for extraction if the preservative solution is not released when the customer snaps the cap onto the funnel during the collection process. A sample could also contain an insufficient amount of DNA simply because there are not enough cells floating around in it. Some people seem to have less DNA in their spit, though almost everyone has enough for the purposes of our analysis.

Step Three: Genotyping

Samples that yield sufficient quantities of DNA are submitted for genotyping,  which scans the genetic material at the approximately 580,000 markers (SNPs) included in our service. This process is also performed in batches of 96 samples. And once again, we monitor the process to make sure it is going as expected. If it isn’t, we try again. In this case a sample is re-run if the analysis process cannot determine its genotype at a minimum of 98.5% of the locations we probe (roughly 570,000 out of the 580,000). If the second attempt fails as well … we ship the customer another sample collection kit for free.

Once a sample has been successfully analyzed, the laboratory sends the resulting data to 23andMe along with the barcode that came with the sample. That allows our database to deposit the information in the proper account and send an automated email notification to the customer who holds it.

With all the things that can go wrong, the sample analysis process may seem a little precarious. But it’s actually quite robust — the vast majority of samples sail right through without a hitch. If things do go smoothly, we can generally return data within eight to 10 weeks of the date a sample is sent to the lab.

My name is Varsha Baichwal and I work here at 23andMe as Director of Operations. You can share your genome with me if you wish – my 23andMe ID is “varsha”.






  • danielgilbert

    An astonishing THIRTEEN weeks after sending you my sample, I get this email: “Our laboratory has received your saliva sample, but unfortunately they were not able to analyze it sufficiently to meet our standards… If you send your sample in as soon as you receive your replacement kit you should be able to see your genetic information in 6-8 weeks.”

    So let me see if I have this right. It takes twice as long for 23andMe to get around to analyzing a sample as they advertise, and then when you need a replacement sample, your customer goes to the back of the line and waits another 6-8 weeks (or is it actually another 13)?

  • http://www.23andme.com MattC

    Our apologies. Processing did take longer than our standard turnaround time due to a sample backlog, combined with the fact that when a sample fails analysis it goes through multiple DNA extraction and genotyping attempts. Our turnaround times are now improving thanks to the measures we have taken to improve the capacity at our contracted lab, and we expect his second sample to be processed within our standard time frame.

  • Londi

    Are there any medical conditions that would cause a person’s saliva to have insufficient DNA? If not, then what else would cause this?

    • Christine

      I have read that fluoride can cause DNA damage, and iodine displaces fluoride, so I wonder if someone who is deficient in iodine may have lower levels of DNA in their saliva? Also, heavy metals such as mercury can damage DNA, so maybe people with heavy metal toxicity may have low levels of DNA? Just a thought.

  • http://www.23andme.com ErinC

    @ Londi

    I don’t know of any conditions that cause low DNA in a spit sample. An interesting theory one of our scientists suggested is that some people have a low pH in their mouths, and this overwhelms the buffer included in the spit kit and destabilizes the DNA. But we don’t have any evidence for this.

    One of the most common causes of not having enough DNA in a sample for analysis is not providing a large enough sample. So be sure to spit until you reach the fill line!

    Here’s some links to the manufacturer of the current 23andMe spit kits if you want to read more about the technical details:

    http://blog.dnagenotek.com/blogdnagenotekcom/bid/35944/Rinse-Swab-or-Spit-What-s-the-Real-Source-of-DNA-in-Saliva

    http://www.dnagenotek.com/DNA_Genotek_Support_FAQs_DNA.html

  • Joseph Becker

    You ducked the question.

    Exactly why does it take 6-8 WEEKS to process the sample?

    Standard industry practice is that a chip is processed in 5-7 DAYS.

    So every thing you described could be done in a 2 weeks max

    -jj

    • http://23andme.com Shwu

      Hi JJ,

      The process could certainly take a week or so for one sample in theory. The 6-8 week estimate allows for scenarios where a sample may fail at one or more steps in the process and may need to be run a second time. It also takes into account the processing capacity at the lab, which is often tasked with analyzing thousands of customer samples in a given week, and additional quality control steps that are taken with each sample. Assuming there were no problems with running a particular sample, we are sometimes able to return data to customers in just a few weeks, but this cannot be guaranteed, as samples do sometimes need to be re-run and there may sometimes be wait times at the lab.

  • Jon

    I think the big thing is people buy the kits, receive them and send them off and then get an email saying it’s been received all in a flurry of activity that happens in a week or so. Then, nothing for months and months.

    Perhaps there could be progress emails, or more accurate ETA emails than just the standard “6-8 weeks” which appears to be an average rather than even a maximum or expected time.

    Surely the lab can provide you with rough timelines and/or feedback on progress at even weekly intervals, saying what stages the genotyping is up to, and this could be relayed onto the clients so they aren’t just fretting and getting irritated and slowly wondering if they made a mistake for months and months…

    • ScottH

      Jon, Thanks for the note and suggestions. We do currently have emails to notify customers when their samples have arrived and when their results are ready. We also recently added a feature that allows people to check the status of their sample. Finally we are literally working overtime at the lab to handle what was a huge influx of orders at the end of last year. We’re increasing our capacity and trying to cut down the wait time. We are making progress toward that goal, although it that doesn’t make it any easier if you are currently waiting for your data. Thank you for your patience.

      • Lynn D

        How long does it normally take to get results?

        Is there a sample chart of what results look like?
        someone said they they get ~1 mil. snp results per person,
        tried to look at the Mendal demo sample but didnt see it.

        Thanks
        Lynn D

  • Mike

    I just received the email that my sample was received. I sent it several weeks ago, so I had assumed it would have been received several days after I sent it. I was disappointed to see that I have another 6-8 weeks.

  • Jeanne Muhlestein

    ScottH, just read your comment about the delay in receiving results from samples submitted to the lab. I am wondering how long the wait period is now?

    • http://23andme.com Shwu

      Hi Jeanne,

      Current sample processing time is approximately 6-8 weeks but this is subject to change as we work to improve this. If you’ve already submitted your sample you should be able to see the status of your sample in your account.

  • Suzanne Pratt

    My husband’s sample could not be processed. We are now waiting on a replacement kit. Are there any tricks that can be used to increase the DNA in the spit? Should he drink any thing before spitting? Not drink anything? I checked his sample, and he did fill it up to the line.

    • ScottH

      Don’t drink or eat anything for at least 30 minutes beforehand. Also rubbing your cheeks on the outside can help a bit.

      • Suzanne Pratt

        Thanks!

        Suzanne

  • Denise

    I was just informed that my DNA extraction has to go through a second round, while my mother’s seems to be okay…for now. We did ours at the same time and the exact same way.

    I’m just wondering why if they can’t get it from the first try, how can they get from the 2nd round of the same vial?

    Also, what’s the percentage rate of getting DNA from the 2nd round I wonder?

    • ScottH

      Denise,
      There are several reasons why analysis of a sample may not be successful, and we don’t necessarily know why it is more difficult to analyze some samples over others. If necessary, the lab will make multiple attempts at all stages of the process in order to provide results; however, due to biological variability some people simply don’t have a high enough concentration of DNA in their saliva for our technology to process.

      If your sample has to repeat one of the analysis steps, your sample status will update; both DNA isolation (Step 3) and DNA analysis (Step 4) may be repeated.

      I don’t have a percentage for you, but it happens. Sometimes we after trying again several times at the lab, we may have to send you a second kit and ask for another sample.

  • Denise

    Thanks Scott,

    So you all run it again several times, not just twice, before sending another kit? Won’t this take weeks?

    • ScottH

      Yes it may take a few weeks to rerun the sample.

Return to top