The live streaming from the MyHeritage conference in Amsterdam this past weekend was both informative and enjoyable. I was even able to watch without getting up in the middle of the night as they stayed available afterwards. Many of the talks included information that was not new to me, although I was pleased to finally be able to watch Leah Larkin’s WATO presentation.
Yaniv Erlich
The talk that impressed me the most was the one by Yaniv Erlich in the middle of day 2 about the DNA health results that MyHeritage is now providing (on a reduced price sale just through Tuesday, Sept 10). With the acquisition of Promethease and SNPedia they have greatly expanded their access to medical genetic knowledge. SNPedia is always my go to resource for looking up specific genes, for example click here to see what it says about the BRCA genes. While Promethease is where you can upload your raw DNA from wherever you tested to get health results (sadly written in formal medical talk) based on current but not always solid research (see comments below) and is linked to SNPedia.
One of the important take aways for me was that if MyHeritage finds a bad mutation in your test, they will use a different method, Sanger sequencing, to confirm the result! This almost completely eliminates false positives. Still if you are diagnosed with a really bad variation, my opinion is that you should double check even further by getting a doctor ordered test like the one from Color Genomics.
Yaniv also emphasized that they have worked hard to present the results in an understandable format. There are some examples of this on their blog (click here).
Another item of importance from his talk was that MyHeritage tests 15 different deleterious mutations on the BRCA genes as opposed to only three tested at 23andme. Those genes are critical for the immune system’s ability to kill cancer cells; thus are particularly implicated in breast cancer among others. Certain population groups, like Ashkenazi jews, have a higher frequency of these mutations.
Testing for the likelihood of getting heart disease and type 2 diabetes were examples used in his talk about another feature offered in these health tests: the use of many SNPs to accumulate a risk score. This is known as polygenic testing. Anywhere from 300 – 40,000 SNPs are used to predict your risk. The interesting point in his presentation was that if you have more of the genetically good variants, you can have a slightly higher BMI without worrying. While, for example, if you are unlucky in your DNA inheritance, you can have a higher risk for heart disease than a smoker!
Polygenic testing statistics slide from Yaniv’s talk
All in all I am impressed. I had dismissed MyHeritage‘s entry into the DNA health field with a “So what, I have already tested at 23andme.” I was wrong to think that. If you have any interest in knowing your risk of heart disease (and are European) or cancer then maybe this test is for you.
I am an affiliate of many DNA companies, and now that includes MyHeritage. At the time I wrote this I was not so I got no financial benefit from this review. However that has changed so that now clicking on their name here or in my footer will give me a small piece of any purchase you make with them.
If someone has a family history of cancer, heart disease or anything, MyHeritage ISN’T the route to take. For example, even though MyHeritage tests for 15 BRCA1/BRCA2 mutations, there are actually in the area of 1000 known deleterious such mutations. So a negative result doesn’t mean that the person doesn’t have a deleterious mutation. A genetic counselor can do medical testing to ensure the right genes are tested for (BRCA1 and BRCA2 are two of multiple genes in which mutations have a cancer association) and that as many techniques for finding deleterious mutations are used. They can also look at family history to potentially identify fact-of a family mutation that isn’t yet detected by genetic testing to give a patient a personal risk to allow for decision making and potential increased surveillance.
As someone whose BRCA mutation was found with 23AndMe, I think tests like these are invaluable for detecting mutations in someone who wouldn’t have thought to be tested otherwise. But they should not be the default for anyone with a known medical history.
I’m also concerned about the acquisition of Promethease. Just in the BRCA arena, Promethease tags many people as having BRCA mutations. But they are mutations that aren’t deleterious–basically, no medical genetic testing would flag the taker as having an elevated risk of cancer. I see people in BRCA facebook groups on a regular basis who are freaking out because of Promethease results. I always encourage them to get medical genetic testing–and so far about 95% have come back with negative results. Promethease is only as good as the data in SNPedia, and that data can be questionable in many areas.
Kitty, thank you for providing the BRAC information.
Lara, and thank you for your insightful comments.
Lara —
Good points. I can’t locate the source right now, but I have a recollection of Yaniv Erlich saying physicians in the US review a family medical history prior to taking the test and recommend expanded medical testing if warranted. The original MH blog alluded to this.
https://blog.myheritage.com/2019/05/introducing-the-myheritage-dna-health-ancestry-test/
I think 23andMe may actually check some of the BRCA mutations that MH does, but they do not have FDA approval for using them in a report.
The highlight I think of all the medical info is that it can open the door to a conversation with your healthcare provider. Family history alone is not enough to get some healthcare providers to initiate an early screening or work up.
And don’t even get me started on adoptees without family history, totally not the place here.
Have that conversation. If your doctor dismisses the findings, ask well what would be the trigger to work up for something, or ask to discuss modifiable risk factors, or ask what can be assessed or tested non-invasively to be sure—or get a second opinion from a specialist.
Preventative & proactive medicine is better than dismissive & reactive medicine.
Anyone interested in medical genetics, I highly suggest doing reading on Platelet Plavix (clopidogrel) Response.
Its a huge issue amongst cardiology.
Excellent write up Kitty, your honesty is commendable!
My colleague Lara makes some very good points based on her own personal health odyssey. My own leads me to some slightly different conclusions.
I am very encouraged that MyHeritage is now emphasizing poly-genetic test results rather than relying on the results of one — or a few — gene(s) — to diagnose physical conditions. I believe poly-genetic testing is the future as we all learn more about the relationship between genotypes and and phenotypes.
While I agree with Lara that we need to consult with genetic counselors and other medical specialists, we need to maintain our own vigilance about our genetic results. I have a genetic mutation that was uncovered as the result of medical research testing on the remnants of a blood sample that was drawn for other purposes early in 2012. Late in 2016 the preliminary result was confirmed by Sanger testing. However, it was another another year and a half before the research team notified my general practitioner of this finding. She was more than willing for me to consult directly with the genetic counselor and with the MD who heads the Vanderbilt Hereditary Cancer Clinic and whose specialty is Genetic Medicine. When I met with the latter, between us we were able to interpret the lab report about the location of the mutant. She understood part of the report which I did not, but there were a couple of items that I was familiar with because of my genetic genealogy work about which she was not familiar. Between us we got through the lab finding.
My point is that both the genetic genealogy community and the medical community — including medical geneticists and genetic counselors are all experiencing a very steep learning curve and will continue to do so for the foreseeable future. We need to work together for the benefit of our health and that of our families.
Fortunately in my case, my phenotype testing has shown no correlation with my genetic anomaly. This seems to surprise my medical team because the medical literature seems to show a close correlation. It raises the issue of “correlation with” being different than being a “causation” when a mutation is found.
We all have so much more to learn and we all need all the information we can get.
Hi Kitty
Does an upload get this info or do you actually have to provide a new DNA sample ?
Thanks!
Cindy
Cindy, an upload is not adequate. They need your actual DNA sample as they are testing many SNPs not tested elsewhere
Cindy, I don’t see a link click on your sidebar to purchase a MyHeritage kit,
If we buy through you do you receive a benefit? If so how do I buy through your blog? Or I will buy you a glass of Pinot? Come to San Francisco and I will share a few bottles with you,…Linda
Linda, Surely you mean Kitty not Cindy?
I get nothing when you buy a kit from MyHeritage only if you buy a 23andme kit through me, this is the link https://www.myheritage.com/health
The words ” DNA health results that MyHeritage is now providing” are linked to that. Many words are linked like “Sanger sequencing” but the link indicator in my posts is a very subtle color change…
Thanks, Lara!!
Promethease and cancer is a cautionary story. They indexed lots of research papers, without any curation of research quality, despite the fact that the early “cancer SNP” research suffered from a reproducibility crisis of epic proportions. Most of the published associations of the SNPs with cancers, until the 2010s, didn’t pan out at all. To give you one example, Promethease reports that rs144848 in the BRCA2 gene is linked with cancer, based on a flawed and long-refuted Chinese report. In fact this genetic variation is present in about 3/4 of people! It’s like, almost everyone is told to worry about elevated cancer risk due to just this one mis-interpreted SNP. In recent years Lara and I made a very strong effort to push Promethease to clean its database from the dangerously sloppy publications of this sort, but they stonewalled us. They say that the Promethease report is a purely research tool and its only goal is to give you a list of papers to read and make your own mind. NOT to make health conclusions. And now MyHeritage toes the same line, that the Promethease tool is a mere index-card catalog to a library of research papers, some good and some flawed, but mostly obsolete.
Contrast it with Kitty Cooper’s words, “Promethease is the place to … get health results (sadly written in formal medical talk) based on the latest research” and I hope you see a major problem 🙁
Another word of caution needs to apply to the “15 actionable cancer mutations” and to the health risk reports derived from multiple SNPs (aka polygenic risk scores). Both of these types of results may be extremely helpful for some of us, literally opening eyes of the test-takers who never ever suspected that they have elevated health risks.
But neither tool is appropriate for the people who are already aware of their risks because of the family history of the diseases, because both “15 mutation testing” and “polygenic risk scores” have whopping gaps in sensitivity. They only identify a modest sliver of the genetic risk factors, missing a whole lot of the disease genetics.
For the infamous BRCA1/2 genes, Lara already mentioned that there is a lot more that 15 mutations. The actual statistics from the not-quite-complete US Government database lists over 5,500 distinct mutations in these genes, and more are discovered daily. Nothing even close to the “list of 15”! Our genetic makeup, and our mutations, reflect our ancestry, the ethnic, geographic, and caste origins of our grandparents and greatgrandparents. Being a genealogy company, MyHeritage approached the problem with ancestry in mind, selecting the commonest mutations in several European populations. For one of ancestries (Ashkenazi Jewish), the “list of 15” gives a modest assurance that most of the mutations are caught. For most of the rest, most of the mutations are NOT caught. For many non-European ancestries, hardly anything makes “the list of 15” at all. Wish the company would be more open about disparate sensitivities of its mutation testing in different ethnicities and races!
The ancestry quandary is just as strong for the “health reports” based on polygenic risk scores. These scores give encouraging results in the whites (capturing a minor but significant sliver of the family risk) but hardly work at all in the minorities. Again, MyHeritage tiptoes around the race issue, unfortunately. Besides, for the heart diseases, limited sensitivity of the scores isn’t the only issue (yes, there are hundreds known high-risk heart disease mutations which should be on the radar screen of the people with strong family histories of the heart disease, like myself). The other un-mentioned issue is that the clinical factors (such as cholesterol and lipids, blood pressure, and smoking) are about as good in predicting health problems as the polygenic health reports. I implore you all to do your bloodwork and check-ups, and to mitigate the risks more efficiently than by the gene scores. (In my family line, the mutation in the LDLR gene makes itself “visible” by extreme cholesterol levels, and can be countered by diet, lifestyle, and statin drugs … no gene testing necessary to catch the mutations’ ill effects in time to prevent the heart condition!)
Ugh, sorry for being such a bore, but I see yet another issue, so “part III”. Kitty Cooper is right on target noting that “One of the important take aways for me was that if MyHeritage finds a bad mutation in your test, they will use a different method, Sanger sequencing, to confirm the result! This almost completely eliminates false positives”.
Of course it does nothing to eliminate false negatives, and the SNP chips are notorious for having both false positive and false negatives. Especially when the SNP is an actionable mutation, which occur in the harder-to-read stretches of our chromosomes in the first place! Just like before, I need to underscore that for the people who knew nothing about their familial health risks before the test, the gaps in the test sensitivity (such as the false negatives) may be just fine. But if you already have a reason to worry about the genetic risks, then don’t let the false negatives give you false reassurances.
I appreciate all the erudite and detailed comments. I have modified my closing statement. Obviously if you have known familial risks you should consult with your doctor and do additional testing to these more easily available ones at MyHeritage or 23andme.
I have a family member with an unusual BRCA2 deletion not picked up by these tests but by the doctor ordered Color Genomics test (he has metastatic prostate cancer).
I was still very impressed by the approach MyHeritage is taking to make genetic issues more understandable and by their polygenic risk concept.
Thank you!
An estranged cousin of mine was (hopefully) the last one to die of heart attack in his 50s, in the 2000s already! He was a born-again Christian who broke ties with the old family over his new convictions, and relied on prayer rather than on medical check-ups. He left kids who have a 50:50 chance of inheriting the killer LDLR mutation which isn’t on the present-day ancestry-and-health tests either. I was unable to pass them a warning, but if the kids do their bloodwork in time, they’d hopefully figure it out before it’s too late.
Would you mind softening the Promethease statement as well? From MyHeritage’s own announcement on the company blog:
“Promethease.com is a literature retrieval service. It allows consumers to upload their raw DNA data (from services such as Ancestry.com, 23andMe, and others) and automatically compare it to SNPedia to see relevant scientific findings regarding their genome”
To recap, it is a literature search tool, not a health tool. It furnishes search results by keyword relevance (not the most recent, not manually curated for credibility). Some published links will be true, many won’t …. careful reading and reviewing is a must.
OK Dmitry, I changed the reference to Promethease to be more accurate. Thanks for your commentary!