Machine Learning Used To Create An HIV Vaccine


When we think of machine learning it’s usually in the context of robotics—giving an algorithm a large set of input data in order to train it for a certain task like navigation or understanding your handwriting. But it turns out you can also train a nasty virus to go to sleep and never wake up again. That’s exactly what the Immunity Project has been doing. They believe that they have a viable HIV vaccine and are trying to raise about $25 million to begin human testing.

The vaccine hacks the Human Immunodeficiency Virus itself, forcing it to mutate into a dormant form that will not attack its human carrier. It sounds so simple, but a lot of existing knowledge and procedures, as well as new technology, went into getting this far. Last week we spoke with [Reid Rubsamen, M.D.] about the process, which began by collecting blood samples from a wide range of “Controllers“. Controllers are people who carry HIV but manage to suppress the virus’s progression to AIDS. How do you find these people? That’s another story which Scientific American covered (PDF); the short answer is that thanks to the work of [Bruce D. Walker, M.D.] there was already a database of Controllers available.

The information accumulated by [Walker] then underwent a data crunching exercise. The data set was so enormous that a novel approach was adopted. For the laymen this is described as a spam filter: using computers to look at large sets of email to develop a complex process for sifting real messages out of the noise. The task at hand is to look at the genotype of a Controller and compare it with the epitope— a short chain of proteins—in the virus they carry. The power of machine learning managed to whittle down all the data to a list of the  first six epitopes that have the desired dormant-mutation property. The vaccine consists of a cocktail of these epitopes. It does, however, require some clever delivery tactics to reach the parts of the world where it’s most needed. The vaccine must not require refrigeration nor any special skills to administer.

The vaccine’s production uses existing methods to synthesize the amino acid peptides, which are the epitopes themselves. The packaging, however, is a new concept. [Dr. Rubsamen’s] company, Flow Parma, Inc., is using microspheres to encapsulate the vaccine, which render it shelf-stable and allow it to be administered through a nasal spray. Learn more about the technology behind the production of microspheres from this white paper (PDF).

If the vaccine (which will be produced without profit) passes clinical trials, it could see mass distribution as early as 2017.

The $25M we mentioned earlier is a tall hill to climb, but think of the reward if the vaccine is successful. You can donate directly to help reach this goal. If you’re planning on giving gift cards this year, you can purchase them for many different retailers through Gyft, who is donating 100% of December proceeds to the project.

33 thoughts on “Machine Learning Used To Create An HIV Vaccine

      1. That’s not how it works. The vaccine has no actual HIV virions. That being said, it’s really not clear how it works. Something to do with surface proteins, but it has nothing to do with antigens or antibodies, and they don’t elaborate on what it actually does apart from claiming it produces robust cell-mediated immune responses.

        Apparently machine learning was used to compare people with natural HIV immunity to those who developed AIDS, and discover what factors allow controllers to avoid developing AIDS for extended periods of time.

        There’s jargon associated with this project that doesn’t appear anywhere else, their animal studies haven’t been published in a scientific journal (nor are they available for the public to read), and the literature on their site is incredibly dumbed down. The scientists look legitimate and respected, but the same could be said for many quacks throughout history. Calling their vaccine “[sic] a completely, novel, disruptive, avant garde, off the beaten path approach” doesn’t engender confidence.

        Color me extremely skeptical, not just of this vaccine, but also of this organization. Let’s give them the benefit of the doubt and say it does indeed work as they claim, that it really is a revolutionary approach to treating viral diseases and will eradicate everything from ebola to hep C to influenza to the common cold. Even if that is the case (and I’m not convinced), you can bet they’re going to patent the shit out of this, viciously attack and potential competitors in court, and charge top dollar for this, squeezing every last penny they can out of it, minus a handful of PR-boosting charity cases in Africa.

        This is only going to be a non-profit effort until there’s a profit to be made.

  1. Sounds very interesting. I happen to be very wary of the many snake oil pushers out there so hope this isn’t one of them. As an unfortunate of said virus I will be studying this closely. Please be a success!

  2. So why exactly doesn’t a big pharma player like Novartis or any Government give the $25M ? It’s peanuts anyway for them. The situation now would be private people funding this, and a big pharma player marketing it afterwards. Or does anyone exspect a small startup to bottle up a vaccine and hand it out to billions of people?

    1. When speaking with Dr. Rubsamen it sounded like there are plenty of sources for funding when it comes to manufacturing and distributing a proven vaccine. But the the money for taking this to clinical trials first is not as easy to come by.

  3. I hope it works. So that they will stop prescribing Reverse Transcriptase Inhibitors which not only block HIV from replicating but also the patients cells. Taking regular doses will kill the patient even if the HIV does not.

    1. I’m sorry my friend, but you’re somewhat mis-informed. Although it is true that RTI’s inhibit replicating cells in general, the have a much much higher affinity for Reverse Transcriptase (which is an HIV protein) than for human DNA polymerase. Thus, very few cells are affected in comparison to the effect on HIV replication. The end result is that regular doses of RTI, although in some cases might make the patient feel worse for wear, *do not* kill the patient.

      -Dr. Z. , M.D.

      1. The toxicity of azidothymidine (AZT) in the treatment of patients with AIDS and AIDS-related complex. A double-blind, placebo-controlled trial.

        “Twenty-one percent of AZT recipients and 4 percent of placebo recipients required multiple red-cell transfusions”

        AIDSinfo Drug Database
        Zidovudine Other Names: azidothymidine, AZT, ZDV

        “Zidovudine can cause serious, life-threatening side effects. These include lactic acidosis (buildup of acid in the blood), liver problems, and blood disorders, including severe anemia. Use of zidovudine for a long time can cause muscle weakness (myopathy).”

        Sounds like AZT a Reverse Transcriptase Inhibitor, A drug that causes anemia so severe the patient needs multiple transfusions to live, has the potential to kill a patient.

        1. Eating too many carrots also has the potential to kill a patient. Dosage and frequency matter a lot. The purpose of a lot of trials and studies is to study the current state of the art and develop the knowledge further, optimise dosages etc.

  4. This looks like a scam, if you visit the site looking for the results they ve had until now, you just get a brief timeline. If they have had any results, they don’t make them public (other than “trials were succesfull”), which is weird because they re a non profit organisation, and they could use the help of others. flashy site, no content.

  5. Very interesting. I tried to contribute them but the form on the keeps responding: “Sorry, your transaction could not be processed at this time. Please try again later.”. Does anybody know if this is a temporary problem or an error?

    1. Hm, thing about AIDS as a control for overpopulation, is the massive amount of death it causes to living people. People just don’t *like* that, for some illogical reason.

      In the meantime you carry on feeling nice and safe, while irresponsible people have sex with each other. Death’s the least they deserve, really.

    2. Condoms are around 80% effective. People like me made very responsible decisions and were unlucky. I’m 24. Thankfully medicine allows me the prospect of a long healthy life. People like you though seem to want people like me to die because you believe the world has too many people. Says a lot about your character tbh. I’d rather have the HIV than be like you :)

  6. Sorry, I don’t buy this. Neither the site looks in ANY way scientificly (does this word exist yet?) serious enough nor the actual information given here in the article or via the videos I could watch.
    To me it looks like a scam to make money by playing around with biochemical equipment.

    I really wish it was true, but it takes a LOT more to make me believe into a vaccine that tells SPECIFIC cells to go asleep and is variable and flexible enough to do this with ALL versions of HIV that are already present (taking the high mutation rate that bummer can come up with into consideration).

  7. Perhaps this is the beginning of the extinction of mankind.

    Ever see the Will Smith movie, “I Am Legend” ? A “cure” for
    cancer turned out to be the instrument of destruction (and
    that movie was just a remake of the classic sci-fi movie with
    the late Charlton Heston, “The Omega Man”).

  8. hmm they first start off that they want to make people like “controllers”
    then what they do is just described as making a vaccine in a similar way that you’d make other vaccines

    sounds more like they wanted to do gene therapy but instead worked on a vaccine

    i think it won’t work on some people because of the mutation rate of HIV

  9. So this dormant form mutation process is irreversible? Is the dormant form still transferable/contagious?

    This work sound more like a cure than a vaccine. It treats symptoms, but doesn’t prevent infection or transmission.

  10. If it actually does that, just think of what could come of it. Much of what we are comes from viruses and bacteria that no longer attack us but still live in us. If HIV is a deadly enemy it might eventually make an excellent friend. Just imagine if it were mutated to destroy cancer cells instead of white blood cells…just an example of course.

  11. No way will big pharma allow a “cure” for HIV (i.e. take this one easy course of drugs then its gone for good) when the current treatments (that you need to keep taking and taking and taking) are so profitable.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.