Wednesday, September 23, 2015

Can dark matter cause cancer?

Image Credits: Agnis Schmidt-May
Tl;dr: Yes. But it’s exceedingly unlikely.

Yesterday, a new paper appeared on the arxiv, provocatively titled “Dark matter as a cancer hazard.” It is a comment on an earlier paper by Freese and Savage, which I previously wrote about here.

Freese and Savage in their 2012 paper estimated the interaction rate of dark matter with the human body for weakly interacting massive particls (WIMPs). They came to the conclusion that the risk of getting cancer from damage caused by dark matter to the genetic code is much smaller than the risk posed by the cosmic radiation we are constantly exposed to.

Yes, dark matter can cause cancer. That’s because literally everything can cause cancer: The probability that a particle collision breaks a molecular bond is never strictly speaking zero, and such damage can potentially turn a cell into a cancerous reproduction machine. Even doing nothing at all can cause cancer, just because a bond may break simply due to quantum fluctuations. It’s not fair, I know. It’s also so unlikely to happen that it didn’t even make it onto the Daily Mail’s List of Things That Can Give You Cancer. Should dark matter go onto the list? After all, the idea that dark matter may lead to “biological phenomena having sometimes fatal late effects” dates back at least to 1990.

In the new paper the authors estimate the interaction probability with the human body for a different type of dark matter. They looked specifically at mirror dark matter whereas Freese and Savage had looked at one of the presently most popular dark matter models, the WIMPs. I can see a whole industry growing out of this.

But what is mirror dark matter and why have you never heard of it?

Mirror dark matter is a complex type of dark matter, a complete copy of the standard model that describes our normal matter. The mirror dark matter interacts only gravitationally with us, or at least only very weakly. This sounds like a nice idea, the next best thing you may think of after just having a single particle. The problem is that we know dark matter does not behave just like normal matter, which renders mirror dark matter immediately implausible.

To begin with there is more dark matter in the universe than normal matter. But more importantly, observations tell us that dark matter must be weakly interacting with itself, otherwise the cosmic microwave background would not have the observed spectrum of temperature fluctuations. Our normal matter interacts much too strongly with itself to achieve that. Then there are case studies like the Bullet Cluster, whose gravitational lensing images reveal that dark matter does not have as much friction among itself as normal dark matter. Dark matter also doesn’t form galaxies in the same way that normal matter does, but rather it acts as a seed for our galaxies. If it didn’t, structure formation wouldn’t come out correctly.

So clearly, dark matter that just does the same as normal matter doesn’t work. On the other hand, a copy of the standard model is a large set of particles with many emergent parameters (like particle abundances) that allow a lot of freedom to make the model fit the data.

You can try for example to adapt the mirror matter model by making changes to the initial conditions, so that they differ from the initial condition of normal matter. The mirror dark matter is assumed to start in the early universe from a specifically chosen configuration, that in particular implies that the two types of matter do not have the same temperature later on. This can solve some problems and make mirror dark matter fit many of our observations. It brings up the question though Why these initial conditions?

As has been argued, probably most vocally by Paul Davies, the distinction between initial conditions and evolution laws is fuzzy. If you fabricate your initial conditions smartly enough, you can make pretty much any model fit the data. (You can take the state that we observe today and evolve it backwards time. Then pick whatever you get as initial state. Voila.) So I don’t actually doubt that it is possible to explain the observations with mirror dark matter. But cherry picking initial conditions doesn’t seem very convincing to me.

In any case, leaving aside that mirror dark matter is not particularly popular because dark matter just doesn’t seem to behave anything like normal matter, it’s a model, and it has equations and so on, and now you can go and calculate things.

To estimate the cancer risk from mirror dark matter, the authors assume that the mirror dark matter forms atoms, which can bind together to “mirror micrometeorites” that contain about 1015 mirror atoms. They then estimate the energy deposited by the mirror micrometeorites in the human body and find that they can leave behind more energy than weakly interacting single-particle dark matter. These mirror-objects can thus damage multiple bonds on their path. The reason is basically that they are larger.

So how likely is mirror dark matter to give you cancer? Well, unfortunately in the paper they only estimate the energy deposited by the micrometeorites, but not the probability for these objects to hit you to begin with. I wrote an email to one of the authors and inquired if there is an estimate for the flux of such objects through earth, but apparently there is none. But one thing we know about dark matter is how much there has to be of it in total. So if dark matter is clumped to pieces larger than WIMPs this means that there must be fewer of these pieces. In other words, the flux of the mirror nuclei relative to that of WIMPs should be lower. Without a concrete model though, one really can’t say anything more.

In the new paper, the authors further speculate that dark matter may account for some types of cancer
“We can thus speculate that the mirror micrometeorite, when interacting with the DNA molecules, can lead to multiple simultaneous mutations and cause disease. For instance, there is an evidence that individual malignant cancer cells in human tumors contain thousands of random mutations and that normal mutation rates are insufficient to account for these multiple mutations found in human cancers [...]”
Whatever the risk of getting cancer from dark matter however, it probably hasn’t changed much for the last billion years or so. One could then try to turn the argument around and argue that if there were too many of such mirror micrometeorites then the dinosaurs would have died from cancer, or something like that. I am not very excited about such biological constraints, the uncertainties are much too large in this area. You almost certainly get better accuracy looking at traces in minerals or actual particle detectors.

In summary, the paper doesn’t estimate the cancer risk for an unconfirmed model. And in any case, short of moving to the center of the Earth there isn’t anything you could do about it anyway.


  1. The universe, and the things that make it up influence us. Dark matter might, but much more likely, high energy cosmic rays, cause genetic mutation. Sometimes this results in cancer. But sometimes this results in a genetic advance. This is a part of what we call evolution. As a scientist, to bring up the "C" word in this context is ultimately a means of self promotion and public sensationalism.

  2. "dark matter must be weakly interacting with itself"

    I think that there is some confusion when dark matter in the form of WIMPs is discussed in the popular literature. Does "weakly" refer to the weak nuclear force, or is it just a generic term?

  3. "Tl;dr: Yes. But it’s exceedingly unlikely."

    Iron-Maiden singer Bruce Dickinson blames his (recently treated and presumably cured) throat cancer on oral sex. No-one should get cancer, but if one has to get it, this is probably the best way. :-)

  4. Phillip,

    In the sentence you quote it's a generic term. In WIMP it refers to a weak interaction, in the sense of a small coupling constant, normally it's *the* weak interaction, but presumably it could be some other interaction. The reason it's a generic term here is (as I alluded to in the recent post about macro dark matter) that there are other ways to get a weak self-interaction than actually having a weak coupling. You could also have an unlikely interaction probability with a stronger coupling, which has a similar result. Best,


  5. Right, which is my point. People read about WIMPs, and figure out that the W is for "weakly" in the context of the weak nuclear force. But "weakly interacting dark matter" doesn't necessary refer to the weak interaction.

    "Dark matter" is also not really clear enough (pun, as always, intended); "transparent matter" would be a better term for what is normally meant, namely unknown non-baryonic matter, which is the "missing matter". In some sense, neutrinos are dark matter, cold gas is dark matter, brown dwarfs are dark matter, that is, they are not visible via emitted electromagnetic radiation, but this is normally not what people mean by "dark matter" in the context of cosmology.

  6. Phillip,

    Yes, I agree with you. The terminology is somewhat misleading, but I don't think it's too terrible. Now consider instead the case of 'dark stars' that are actually bright... Best,


  7. Xenon100, LUX, ZEPLIN-III, PANDA-X, DEAP, ArDM, WARP, DarkSide; CDMS, CRESST, EDELWEISS, EURECA; SIMPLE, PICASSO saw nothing. Fermi-LAT, DAMA/NaI, DAMA/LIBRA, CRESST saw near nothing, then otherwise explained. People are not better collision detectors.

    If you fabricate your initial conditions smartly enough, you can make pretty much any model fit the data. Test the obvious first. Baryogenesis requires vacuum trace chiral anisotropy toward hadronic matter. Physically and chemically identical single crystal test masses in enantiomorphic space groups (opposite shoes) in a vacuum left foot locally vacuum free fall non-identically. Space groups P3(1)21 versus P3(2)21 alpha-quartz,
    Geometric Eötvös experiment, 6.68×10^22 pairs of 0.113 nm^3 opposite shoes.


Comment moderation on this blog is turned on.
Submitted comments will only appear after manual approval, which can take up to 24 hours.