I recently picked up an AncestryDNA kit out of curiosity to find out how well the data matched up to the 23andMe test that I did a few years ago. Quick answer: It matched up better than I expected.
First, a couple of caveats:
I’m not a genealogy expert and was not comparing the two tests as far as accuracy of determining my ancestry. I’m also not a statistician, so the mathematical comparisons of the raw data files are just the basics.
Taking the test:
Both companies are fairly similar in the simplicity of getting the testing done. You order the kit — either through the company websites or through Amazon.com — and it comes in the mail. The box contains a vial to spit into, instructions on how to register the kit, and a small pre-paid shipping box to mail the vial back to the company.
Both 23andMe and Ancestry.com advertise that it takes 4 – 6 weeks to get the test results back after they receive your vial of spit. It was faster than advertised (about 2 weeks for AncestryDNA) when I did the tests, but I think the times can vary depending on how busy the lab is when you send in your test.
The privacy policies:
Once you have taken the test, you also have the option of answering research survey questions on 23andMe and on Ancestry.com. Be sure that you understand that you are giving your survey information to the companies to use for their own purposes.
Downloading the raw data:
Both companies allow you to download and keep your raw data file. I highly recommend that you do so as soon as you get the results. The information is yours, and you should keep it safe.
Here are directions on how to download the raw data:
Download your genetic data from 23andMe.com
Download your genetic data from AncestryDNA
Both companies also have a clearly stated way to delete your data from their records if you choose to close your account with them. Here are the directions: Deleting your 23andMe account; for AncestryDNA, there is a button to delete data right under the download link on your settings page.
Searching your raw data online:
23andMe.com has a convenient interface for searching through your raw data on their website. It is in their Tools section, under Raw Data. You can search by rs id number or by gene name. AncestryDNA does not seem to have this option.
Using your raw data file:
The raw data file for both companies comes as a zipped text file. Both files include the rs id #, chromosome, position, and your genotype. AncestryDNA’s data is formatted a little bit differently in that the genotype is given separately as “allele 1” and “allele 2”, where 23andMe combines the information into a “genotype” column.
You can simply open up the text file on your computer and do a “Find” to search for a rs id number. Everyone should have the ability to open a text file on their computer, no matter the operating system.
A better option (in my opinion) is to import the text file into Excel. To do so, open a new Excel Workbook and click on the Data tab. There should be an icon there labeled “Text” that will let you import a text file. Both the 23andMe and AncestryDNA files are tab delimited. Simply accept all of the default setting in Excel for the text import.
Importing it into Excel then gives you the option of using a second worksheet to make notes on what you learn from your genetic data.
Comparing the raw data files:
I decided to compare my 23andMe (v. 4) data file with the AncestryDNA file. 23andMe gives data for over 600,000 nucleotide base pairs, and AncestryDNA’s raw data covers over 650,00 base pairs. Comparing the two files, there were over 303,000 rs ids in common between the two. (This isn’t a completely accurate comparison since 23andMe reports some of the chromosome positions in a proprietary i-number format instead of as a rs id, but it is close enough for my purposes.)
Of the ~303,000 rs id’s in common, for my data, there were just over 1,000 for which the genotypes did not match. This comes out to 0.3% that did not match — or, alternatively, 99.7% that did match.
Which test is more accurate?
Knowing that for my data the two data files matched for 99.7% of the data actually doesn’t tell me anything as far as which one is ‘correct’ for the ~1,000 genotypes that differed. Neither company guarantees that their testing is accurate, and both companies are very up-front about it with disclaimers stating that it isn’t being offered as a medical test.
I was actually expecting the mismatch percentage to be higher between the two tests. While I’m not an expert on error rates in genetic sequencing, several studies that I had read lead me to expect that there would be more variation in the tests.
Everyone who is doing either AncestryDNA testing or 23andMe testing needs to read the privacy policies and also understand that the data shouldn’t be used as the only basis for making major medical decisions. I’m fine with a little uncertainty in looking at my genetic data for something like deciding that I should eat more foods that are high in choline or add in more leafy greens for folate. Any major health decisions should always be double checked with a test ordered through a lab certified for that test.
Enjoy this comparison and planning on buying a test kit? I would appreciate you using my referrer link, which will cost you nothing but help me keep on blogging.