An interesting conversation and a new article.
Hi there,
I would like to share a conversation that I had recently. For me, it was one of those eye opening conversations that has stuck with me.
A few weeks ago, I made a spur-of-the-moment trip to a genealogy conference sponsored by AncestryDNA, 23andMe, and MyHeritage, with the thought of having a booth there next year.
While in the vendor hall, I talked with lots of people... including two guys in nice suits, sitting at a booth with nothing but their laptops. No company banners, no free pens, nothing that would be expected for a conference booth.
Intrigued, I asked them what their company does.
They told me about their startup company for 'AI' software to identify if someone in a photo is related to you. Scan in a photo of yourself and then scan in a group photo of unknown people; the software should pick out the person in the group photo who is your cousin. Their idea was that you could scan old photos and see if the person was your great-grampa, or something like that.
Nifty concept, and I kept asking questions...
It turns out that they were at the convention to make contacts with MyHeritage, Ancestry, etc. in hopes of integrating with their systems. They also mentioned that their "AI" tool used both phenotype and genotype information.
When I asked a few more questions about other uses for the software, it became clear that the startup was an offshoot of a bigger company. Real time surveillance imaging for threat detection from public cameras was part of it. At this point, they cut off the conversation to go talk with someone else.
As I walked away, I couldn't help but think about the implications of their work.
For machine learning software to work, it needs large data sets to train on. While I didn't get the chance to ask them directly, I'm assuming that a fun little AI tool to identify your grampa in a photo wasn't the real goal. They need hundreds of thousands of photos of people with known relationships (e.g. family trees with pictures) to train their machine learning algorithm to spot familial connections. While it will likely be a fun app to use on old family pics, the trained machine learning algorithm to detect relationships will have other potential uses. Say that your cousin is on a terrorist watch list. If you're walking down a public street or through the airport, this type of software could pick you out of the crowd and label you as a close relative of the terrorist.
What's my point with this rambling story? Be thoughtful about where you upload your genetic data and photos. I have no idea whether the 'startup' company for AI family photo recognition will ever get off the ground or if they were able to connect with the big ancestry companies. However, the potential is there for someone to make the data connections even using public family trees. This reminded me of a news story from a couple months ago about a lady who was kicked out of Radio City Music Hall. She was there with her daughter's Girl Scout troop to see the Christmas show. However, security pulled her out of the group because their facial recognition system identified her as working for a big law firm that was in some kind of litigation with another Madison Square Gardens property. The facial recognition software was trained to detect all the law firm employees based on the hundreds of profile photos on the firm's website. Just something to think about.
Gratefully yours,
~ Debbie
New article!
Lymphedema: Causes and Genetic Pathways
Key takeaways:
~ Lymphedema is caused by interstitial fluid building up under your skin, often in the legs or arms.
~ Impairments to the lymphatic vessels prevent the fluid from moving out of the tissue.
~ Genetic variants can increase the risk of lymphedema, pointing you towards the underlying cause as well as targeted solutions.
Monthly Giveaway!
Enter to win a free annual membership!
What I've been reading:
1. The Humans Who Live As Lab Rats (Free Press)
This Free Press article is an interesting look at the people who are making a living by participating in clinical trials.
Randomized, placebo-controlled clinical trials are the gold standard for understanding whether a medication is safe and effective. While I'm aware of the shenanigans that can go on with statistical analyses in these trials, I had never thought much about the people who participated. The article explains that some people participate in lots of clinical drug trials, making essentially a full-time wage for it. My first thought, though, is that a 'professional' clinical trial participant isn't really the random, totally healthy person that I assumed participated in studies.
2. Telehealth startup Cerebral shared millions of patients’ data with advertisers (TechCrunch)
From the article: "Cerebral has revealed it shared the private health information, including mental health assessments, of more than 3.1 million patients in the United States with advertisers and social media giants like Facebook, Google and TikTok.
The telehealth startup, which exploded in popularity during the COVID-19 pandemic after rolling lockdowns and a surge in online-only virtual health services, disclosed the security lapse in a filing with the federal government that it shared patients’ personal and health information who used the app to search for therapy or other mental health care services."
My two cents: Keeping your health concerns private is why I don't have ads on Genetic Lifehacks. This is also why I don't embed social share icons on each page. Facebook doesn't need to know that you are reading about Alzheimer's genes or prostate cancer genes.
3. The association between light exposure before bedtime in pregnancy and the risk of developing gestational diabetes mellitus
New study on how light in the evening (e.g. phones, tablets, TVs, bright overhead lights) increases the risk of gestational diabetes. This isn't surprising. Light at night suppresses melatonin production. Melatonin receptors on the pancreas help to regulate insulin and overnight blood glucose levels.
This study did not include genetic variants, but there is a melatonin receptor variant linked to an increased risk of diabetes in people who eat dinner later in the evening. Members can check their MTNR1B gene here.