User:Timothy Richard Gall
|Summary||Born in 1979; sex is in the report|
Since this wiki could be read by anyone: 23andMe is actually fun to use. only use promethease if you want to learn some genetics
But for the geneticists out there:
If you would like the full data of me deCodeMe and 23andMe downloads, please ask. I don't know if there is any public place to post them. If anyone knows of one, please email me at (fill in the dots) Timothy Gall at US Army (mil). My medical records are already electronic, but I have no standard way of copying them.
- BP from 90/50 to 125/80
- Height from 180 to 182.5 cm.
- Weight from 133 to 155.
- Body fat from 4.5% to 12%
- Eye Color Hazel
- Hair Color Black
- Cholesterol HLD=55, LDL=120
Although my heart attack risk is <1%, I have had family members have heart attacks with similar cholesterols and BPs. I weigh significantly less than them, though.
Run times are above the 50% mark for all lengths from 200m to a marathon. I exercise (mostly running) and have a good diet (low fat vegetarian, although I take two fat pills to get extra omega-3 fatty acids in my diet). I don't smoke, but rarely I will have a dry red wine or an oatmeal stout. I should mention that the only family history I know of is multiple sclerosis and heart disease. Several members of my family have had cancer, but there is no familial pattern. I personally don't have any major medical conditions, although I have several minor ones, e.g., astigmatism and spondylosis deformans.
Here are some interesting quirks I have found so far.
1. Both services put information stating that I have very large (both in a relative and absolute sense) risk for several diseases, including basal cell carcinoma, type II diabetes, and psoriasis. They often use different SNPs, though, so the relative risk are often quite different. E.g., deCodeMe put my RR for ARMD at 1.35, and 23andMe put at at .18. The corresponds to a 11% vs a 1% risk. The quirk is that they put only a number, not an interval. While a prediction interval for an individual would probably be useless, a confidence interval could still be informative.
2. I found the information from promethease to be more interesting and useful. Interesting because it shows much more information than the commercial services. Useful because the descriptions of the research, as well as the sheer number of SNPs with very small and uncertain effects give a much more realistic view of the usefulness of the information.
3. Even though both services measure the risk of psoriasis using the same SNPs, and even though they had the same calls for those SNPs, they gave different population, relative, and absolute risks!
Locus Chromosome Variant / SNP My Codes Relative Risk Genotype frequency #Cases / #Controls Study HLA 6 rs10484554 CT 2.67 (1.89) 15.9% 810 / 1256 - Liu Y et al. PLoS Genet. 2008 Mar 28;4(3):e1000041. IL12b 5 rs3212227 AA 1.13 (1.12) 73.4% 1446 / 1432 - Cargill et al Am J Hum Genet. 2007 Feb IL23R 1 rs11209026 GG 1.05 (1.06) 86.5% 1446 / 1432 - Cargill et al Am J Hum Genet. 2007 Feb
Note that a NCBI search for rs10484554 listed only one publication, so 23andMe must have estimated the risk from the same source!
These are the data from deCodeMe, with the relative risk for 23andMe in parenthesis. Due to the HLA-C relative risk being so much higher for deCodeMe, it gave me a total relative risk of 3.17 versus 1.98 for 23andMe. But since deCodeMe put the average lifetime risk for male Caucasians at 2%, and 23andMe put it at 10.7%, 23andMe estimated my absolute risk higher than deCodeMe. In fact, deCodeMe, with a 3.17 RR, put my absolute risk at 6.3%, which was still lower than the average risk of 23andMe. 23andMe put my absolute risk at 21.2%. The following table should make it clearer:
Absolute Risk Relative Risk Population Risk deCodeMe 6.3% 3.17 2% 23andMe 21.2% 1.98 10.7%
4. I know from having several EKGs that I have a long QT interval. This insignificant fact immediately drew my attention to the just as insignificant fact that this may be genetic.
5. I have long known that I cannot taste PTC because in my sixth-grade biology class I tasted it as part of the genetics instruction. I also learned common traits such as widow's peak, hitchhiker's thumb, and attached earlobes were genetic in that class. Sadly, molecular genetics hasn't kept stride with 6th grade biology. Only the ability to taste bitter was mentioned by deCodeMe, 23andMe, and promethease. I did learn that it was slightly more complicated than I was taught, though. It involves three SNPs, not one.