5 ms·

Be careful interpreting the numbers about how much risk is increased for certain genes. For example, let's say they point out that a certain gene increases the

by natch 11y ago

Be careful interpreting the numbers about how much risk is increased for certain genes. For example, let's say they point out that a certain gene increases the risk for some malady by 100%. If the average risk across the entire population being used as a baseline was 0.02% (not 2 percent, but 2 tenths of a percent), then increasing that risk by 100% brings the risk to 0.04%, meaning 4 tenths of a percent, still a very low risk. The point is, the way they present the risk numbers is overly alarming and designed to get you to buy in to the product. I think some people see numbers like 70% or 100% and immediately think that means those genes indicate a 70% or 100% chance of developing the disease. This is far from the case.

vitd 11y ago
Exactly! A friend recently got their genetic data from 23 & me and used Promethease to analyze it and it was full of scary sounding stuff like this. I had to talk them down and explain exactly what you did above.
natch 11y ago
Correction, can no longer edit: 2 and 4 hundredths of a percent, not tenths, oops. To be fair, some of the risk factors are a lot higher, for example say with a risk of 20%, a 100% increase brings the risk up to 40%, which is not something to dismiss, while still not being the scary-sounding 100%.
icegreentea 11y ago
Actually, if you look at how Color has specifically called out most of the risks in their material, they very carefully focus on life-time risks, instead of 'rates'. So, unless they are completely lying with their numbers, you can actually look at their 'risk' rates in the naive way. For example, the BRCA mutations actually result in a 2-3 fold increase in life time risk in breast cancer (in women), resulting in a 50 to 80% (depending on what study you trust) life time risk in breast cancer. (http://www.ncbi.nlm.nih.gov/books/NBK1247/ http://www.ncbi.nlm.nih.gov/books/NBK1247/) It's kind of weird, but also kind of obvious. Color is choosing to focus on only the highest correlation/impact mutations. They also require physician buy in on every test they sell. Because of that, they're really just going to focus on the blockbluster genes, the ones which we actually have evidence of high life-time risk. I mean seriously, their BRCA tests by themselves should sell this test - $250 USD is very affordable as far as genetic tests go. Oh, and hilariously, Color chose to spend some time/space to focus on BRCA (rightfully so, it's one of the great validating cases of the power of genomics and the potential for personalized medicine) without talking about one of the reasons why BRCA tests were kind of stagnant for 20 years was cause BRCA1 and BRCA2 (the genes) were patented, along with any diagnostics involving the those genes.

by natch 11y ago

vitd 11y ago

Exactly! A friend recently got their genetic data from 23 & me and used Promethease to analyze it and it was full of scary sounding stuff like this. I had to talk them down and explain exactly what you did above.

natch 11y ago

Correction, can no longer edit: 2 and 4 hundredths of a percent, not tenths, oops. To be fair, some of the risk factors are a lot higher, for example say with a risk of 20%, a 100% increase brings the risk up to 40%, which is not something to dismiss, while still not being the scary-sounding 100%.

icegreentea 11y ago

Actually, if you look at how Color has specifically called out most of the risks in their material, they very carefully focus on life-time risks, instead of 'rates'. So, unless they are completely lying with their numbers, you can actually look at their 'risk' rates in the naive way. For example, the BRCA mutations actually result in a 2-3 fold increase in life time risk in breast cancer (in women), resulting in a 50 to 80% (depending on what study you trust) life time risk in breast cancer. (http://www.ncbi.nlm.nih.gov/books/NBK1247/ http://www.ncbi.nlm.nih.gov/books/NBK1247/) It's kind of weird, but also kind of obvious. Color is choosing to focus on only the highest correlation/impact mutations. They also require physician buy in on every test they sell. Because of that, they're really just going to focus on the blockbluster genes, the ones which we actually have evidence of high life-time risk. I mean seriously, their BRCA tests by themselves should sell this test - $250 USD is very affordable as far as genetic tests go. Oh, and hilariously, Color chose to spend some time/space to focus on BRCA (rightfully so, it's one of the great validating cases of the power of genomics and the potential for personalized medicine) without talking about one of the reasons why BRCA tests were kind of stagnant for 20 years was cause BRCA1 and BRCA2 (the genes) were patented, along with any diagnostics involving the those genes.