Disease Risk Prediction from AI and Advanced Genomic Analysis

Michigan State University senior vice president Stephen Hsu, a theoretical physicist and the founder of Genomic Prediction discusses how AI and super-cheap human gene sequencing is revealing the secrets of genetics and will enable an explosion of disease cures.

Human gene sequencing prices have been falling a rate many times faster than Moore’s law rate of lowering computing costs.

* FDA approval of the first genetic treatment for monogenic conditions and the work towards developing treatments for polygenic conditions like diabetes and cancer.
* How this technology might exacerbate existing social inequalities or create new ones; is it just an issue of access, or does it go further?
* Developing best practice protocols for governance and regulation of genomic technologies.

Genomic Prediction is a startup company that provides disease risk scores based upon AI and advanced genome sequence analysis.

Biorxiv – Genomic Prediction of Complex Disease Risk

They construct risk predictors using polygenic scores (PGS) computed from common Single Nucleotide Polymorphisms (SNPs) for a number of complex disease conditions, using L1-penalized regression (also known as LASSO) on case-control data from UK Biobank. Among the disease conditions studied are Hypothyroidism, (Resistive) Hypertension, Type 1 and 2 Diabetes, Breast Cancer, Prostate Cancer, Testicular Cancer, Gallstones, Glaucoma, Gout, Atrial Fibrillation, High Cholesterol, Asthma, Basal Cell Carcinoma, Malignant Melanoma, and Heart Attack.

They obtain values for the area under the receiver operating characteristic curves (AUC) in the range ~ 0.58 – 0.71 using SNP data alone. Substantially higher predictor AUCs are obtained when incorporating additional variables such as age and sex. Some SNP predictors alone are sufficient to identify outliers (e.g., in the 99th percentile of PGS) with 3 – 8 times higher risk than typical individuals. Substantial improvements in predictive power are attainable using training sets with larger case populations. There will be rapid improvement in genomic prediction as more case-control data become available for analysis.

Genetic analysis can now identify risk outliers (e.g., with 5 or 10 times normal risk) for about 20 common disease conditions, ranging from diabetes to heart diseases to breast cancer, using inexpensive SNP genotypes.

SOURCES- Infoproc, Biorxiv, Youtube

28 thoughts on “Disease Risk Prediction from AI and Advanced Genomic Analysis”

  1. I’d think a more intelligent person could understand and learn history better than a stupid person.
    Unless you mean they won’t be able to relate to the stupid decisions of the past?
    “But I don’t understand Sir. Why would the people of the 1970s introduce flared polyester trousers in the first place? “

  2. Probably computers not men figuring the odds. And as far as genetic engineering goes with absolute control there will be no tradeoffs. But till that time comes we can always reverse anything we do genetically.
    Problem with intelligence versus fertility? Just reverse the genetic makeup for a few months or days.

  3. When I was younger the great science fiction writers wrote of aliens with living bio ships. I always thought that was dumb and technology would be better. Now I am older I kind of understand more. You mentioned nano technology. If you think about it that is exactly what living things use everyday and make changes to their body with.

  4. You about hit the nail on the head. We are going to go extinct because of climate change. It is to late for us to prevent it(as we are now), no mater what we do. But increase everyone’s intelligence by a factor of 2X or more and we will be able to re-terraform Earth very quickly.

  5. Once again think outside the box. Why would the changes to the brain and head HAVE TO occur before birth? Keep in mind these changes could be done to a adult as well as a embryo.

  6. All of you are exactly right I was using IQ as synonymous with intelligence for convenience but of course they are not synonymous.

  7. Ok, I will give you proof but you are to set in your ways to believe absolute hard facts. You will believe what you want to believe not the facts.
    The romans, greeks and egyptians kept detailed records of inventions and discoveries and when they occured. By plotting those inventions for last 2,000 years, you can get a very reliable estimate of the growth of technology based on hard facts.
    The growth is so hard to believe many scientist and laymen repeated the study with the same results
    The error rate when predicting this growth into the future can be calculated like this — projected years/2000. So a 50 year projection would be 50/2000 = error rate = 0.025
    Ray Kurzweil did the study. Check out who he is on wikipedia.
    Here is the study. He was featured in time magazine and forbes magazine as the second Edison.


  8. People have a weird tendency to look at one technology, and extrapolate its effects out into the far future while assuming that everything else stays the same for some reason.

  9. At some point, you get all the same GDP boosts by letting AI do the thinking for you anyway. Then the intelligence we choose to have is purely a lifestyle/cultural choice. We’d just be using our big heads to make conversation, jokes, and art. Are we happier if we make conversation about Aristotle (in Greek), jokes about Proust (in French), and art that incorporates string theory (in computer code)? Probably not.

    But a little bit of an intelligence boost would be useful right now, and even in the future it would help us avoid acting against our own interests.

  10. rderkis, I’m a little less optimistic but mostly with you here. I completely agree that people should remember that we can always reverse these genetic changes. It’s just so weirdly, pathologically pessimistic to think that we could figure out how to enhance our genomes to become smarter, and then somehow be too stupid to understand how to mess with our genomes again. There is some real danger that we might develop something that harms us, but we enjoy it so much that we just aren’t willing to give it up. But as geniuses, you’d think we’d figure out a workaround.

    But here’s a wrinkle to consider – all technologies become obsolete. And genetics will become an obsolete method of intelligence enhancement.

    We can definitely get 50 or so IQ points with biology alone and in normal size heads. At some point though, it will just be easier to work with chips. Or whatever future nano-thing replaces chips.

  11. I agree that letting our genes be determined by culture instead of natural selection is a whole new ball game, which really could lead to consistently low fertility rates if fertility necessarily involves some trade-off, like intelligence, that we just don’t want to let go of. Over many generations this could lead us into our extinction.

    But here’s the thing – ever since nuclear weapons were industrialized, a few people have had the ability to end all of humanity in a few hours. This genetic technology has the ability to end all of humanity over hundreds of years, if everyone in each generation decides to let it every step of the way. And these would be very smart people making that decision each time.

  12. I’ve heard it said that because of the way IQ scales, every 10 or 15 IQ points is the ability to handle tasks twice as complex or so. So call it ~35 IQ points for people that are 5-6X as smart. If you’re starting from an IQ of 85, that manages to get you through university.

    I don’t see that happening for adults from a single shot with near term technology, but it should soon be possible for embryo selection.

  13. Care to specify what you mean by near future? From where I am standing, you seem to have very little to back up your claim other than conviction…

  14. And you would want that? Bubble headed humans where no baby is born naturally?

    To get that, you would have to engineer the sexual preference as well so that men and women would be A-OK with giant heads. Furthermore, you would either have no woman give birth to a child naturally, or you would have to engineer their hips to be even wider than today as to allow giving birth to even bigger headed babies than today. So either giant hips or baby factories…

    And the neck muscle would have to be strengthened and the spine (as to avoid back pain). Even such an “improved” human would probably not be very athletic, particularly women with the giant hips, so such things as running and playing would be pretty much gone.

    By changing humans you would cut them of from history, both because of their changed anatomy, changed sexual preferences and their increased intelligence. And for what, really? A boost in GDP? Or are you naive enough to believe that this increase in GDP would make future humans happier or more fulfilled?

  15. You are right there are to many variables for us to foresee the consequences, that is why we will be using a quantum computer to handle it.
    Most of you are making the same mistake as R. Kimhi. You must think outside the box a little. In the near future any change we make to the DNA can be reversed and re-reversed at any time. Just like a blueprint.
    Why would you think with complete control and understanding of the DNA, we could not increase intelligence? Now I don’t think we will need to go the route of increasing the brain size but we could go that route if need be. The DNA that regulates bone growth of the cranium would need to be tweaked along with the blood flow etc. Now I am not naive enough to believe it will be that simple but it can be done. All we need to do is increase intelligence a little in everyone and that will start a avalanche of improvement in intelligence.
    I am 72 years old and I can see what’s coming, but I can’t understand with the history of invention why you can’t. How many times has something been done when everyone else said it was impossible?
    Who has successfully stopped progress for more than a few years?

  16. Arthur C. Clarke said, “Any sufficiently advanced technology is indistinguishable from magic.”?

    That will seem trivial compared to how far they will go in the near future. And when I say they, I mean you and everyone else.

  17. Very sensible, it has to be done very carefully. For example some genes are beneficial when they show in one allele or in certain combinations but harmful in two, or in other combinations. There is more than one layer of reading the genetic information and we know very little about that. Also we are not sure what is the role of the junk in DNA to name a few issues. Totally agree about your comment about intelligence. Also a lot of the psychotic traits exist because they are beneficial in certain situations of social stress. By the way this is all true also for genetic editing. But I think that even now we can go for weeding out certain mutation or make sure that they don’t show in both allales.

  18. Two things. First off, we dont really know what side effects such a selection would bring. Perhaps this would whipe out some useful genome variations the make some humans resist certain diseases..? What if the genetic monoculture would die of a new “black death”?

    Second, I’m not sure that selecting for new IQ extremes would make the kids more successful. The average is the average for a reason. We know, for instance, that high IQ women tend to have less children than medium and low IQ women. So a lot of well meaning parents could make sure that their kids would be really intelligent, and as a consequence, make sure that they in turn have few or no kids…

  19. You are exagerating. There is no way to create – even with perfect control and knowledge of the genes – a human with an IQ score of 500-600.

  20. I see a real utility in this test. You take it, and then you adapt your habits to minimize the risk of dying from the illness that is most likely to afflict you. For instance, if you have a great likelyhood of developing malignant melanomia, you stay out of the sun.

    Similarly, you test often for the diseases that you are likely to get, so that you have a good chance of starting treatment early.

  21. A utopian future has begun. Soon due to genetic engineering there’ll be no disease, hunger, pain, old age etc. And when I say soon, I mean soon. There are at least 3 clinical trials in humans using genetic engineering going on now. They are now offering genetic engineering for pet dogs. A dog’s drug/therapy only takes 1 year for FDA approval, so research on dogs is surging forward. Much of that dog research will be applicable with modifications to humans. You can even buy cheap genetic engineering kits on ebay for your own personal experiments. The two little girls in China who had their genes changed so they can resist AIDS also got a boost in their IQ.
    Of course the greatest advancement of all is when they enhance EVERYONE’S intelligence with a single shot/treatment. And when I say enhanced. I am talking a factor of 5X or 6X or more.

  22. The way to go. Instead of gene editing that is going to create synthetic human beings with unexpected unforeseen yet negative consequences we can work with mother nature and carefully weed out negative genetics from the human specie by predicting them even before conception and definitely in the early pregnancy days before the spirit enters the not yet alive embryo. With this knowledge and some incentives couples will make better choice of which children to bring to the world!


Leave a Comment