I did my 23andMe test in July 2016, and at the time, I downloaded and did my own analysis of the SNP raw data using different methods. Yesterday, I decided to check if the raw data had changed since that time. It did. I also looked at what’s new with the reports on the web site and found some had changed too.
For example, new in the 2019 report is that “my genetic muscle composition is common in elite power athletes”. That sounds great but I wish I had known it as a child, then I would have perhaps focused on “professional sports”! Last time I checked the report they described the gene differently, as a gene that reveals whether I am a long-distance runner or a sprinter. It shows that although my gene obviously stayed the same, more information about the gene itself was revealed through new scientific studies. I think “elite power athlete” is more motivating!
New in 2019 is the list of warnings to people who download the raw data that seem to discourage them to get a second opinion from another service:
“If you decide to download your raw genetic data from 23andMe, take caution in uploading your data to any third-party applications or any service outside of your 23andMe account. “
The reason they cite is that the raw data is not validated for accuracy except for parts of it that were used in their report. Analyzing the data with a third-party application can actually help validate the results in their reports. Moreover, any new findings, can then be validated with targeted genetic tests. Also, among the warnings is this one:
“By downloading your raw data, you could discover sensitive health information about yourself or family members with whom you share DNA”.
It is important to not make such sensitive information public, just like with financial and other health information. We all do however share financial information with financial advisors, and similarly, the owner of the health data has several options for their raw data. Using a third-party application like Promethease for a “Second opinion” analysis is a good option, or getting a service from a health data analyst at Omicshelpdesk.com to get personalized analysis and validation of the data.
I checked the 23andMe Changelogs which explain how the raw data has changed over time and saw that between 2016 and now the data had changed once:
“July 27, 2017 As part of our continuous efforts to improve the quality of data present in your raw data download, the number of SNPs available in your download may have changed.”
SNP stands for Single Nucleotide Polymorphism and indicates what base (A,C,G,T) is found in that position in either of a person’s pair of chromosomes.
I compared the two raw data files. I found that some SNPs were removed and some were added from the previous file but the base calls stayed the same. For example, rs2229579 was added, and i6034144 was removed. The text file contains 601,818 rows, 1 for each SNP, that look like these:
rs2239804 6 32411523 TT
rs7192 6 32411646 TT
rs3129888 6 32411726 AG
rs2239802 6 32411846 CG
The first column, the SNP name, can be entered into the search box of https://www.snpedia.com. Sometimes no results appear but for example, for rs7192, the results return a description of a SNP linked to peanut allergy.
The 23andMe raw data is actually easier to analyze than full genome sequence data which has 3 billion of those SNP bases in each chromosome and for which most variants have not been studied yet. But science is advancing rapidly and most genotyping companies like 23andMe will eventually be forced to switch to whole genome sequencing.