Joshua Swamidass is an assistant professor at Washington University’s medical school, an MD and computer science PhD. He’s associated with the theistic evolution advocacy group BioLogos, and at the organization’s Open Forum page, Dr. Swamidass offers an argument against intelligent design from the existence of cancer.
Noting a new article he co-wrote in Nature Genetics (“Unsupervised detection of cancer driver mutations with parsimony-guided learning“), he says he wishes to point out the “usefulness of evolutionary theory. It turns out that evolutionary theory is indispensable to understanding cancer.”
Swamidass highlights a post from a National Geographic blog, The Loom.
Cancer evolves. Those two words may sound strange together. Sure, birds evolve. Bacteria evolve. But cancer? The trouble arises from the fact that cancers, unlike birds and bacteria, are not free-living organisms. They start out as cells inside a person’s body and stay there, until they’re either wiped out or the person dies.*
Yet the same forces that drive the evolution of free-living organisms can also drive cancer cells to become more aggressive and dangerous. Evolution becomes our inner foe if mutations disable a cell’s self-restraint. The cell multiplies. Sometimes a new mutation arises in its descendants. If the mutations allow the cancer to grow faster, the cells carrying it will take over the population of cancerous cells. Natural selection and other processes that drive evolution on the outside start driving it on the inside.
Cancer in human beings has been around for a long time, currently traced back 1.7 million years. In a particular individual, the “evolution” of a sickness, a tumor, doesn’t at first hearing sound like anything relevant to ID, which questions the creative power of evolution to produce complex specified information, not runaway cell multiplication resulting in death or debility for the biological beings that, in fact, evolution needs to explain.
The National Geographic post goes on to cite an example of simple drug resistance that cancer overcame. There’s nothing unusual here. No complex new mechanisms evolved. We read, “Evolution becomes our inner foe if mutations disable a cell’s self-restraint.” That sounds like evolution by breaking things, not building them — “devolution,” as Michael Behe puts it in the context of antibiotic-resistant bacteria, not “evolution.”
Surely there’s more to Swamidass’s case than that.
A jocular BioLogos reader crows:
This is fantastic. Evolution vindicated yet again with actual real world experimentation and testing. I wonder what the ID explanation for cancer is? Do they think it was “intelligently designed”?
I think they would correctly point out that cancer is not a new species.
Well, yes. The origin of species is what evolutionary theory seeks to explain. Anything else?
The bigger question is this: if (1) evolutionary genetic tools correctly infer the progress and history of cancer, (2) cancer regularly innovates with proteins of novel function, (3) regularly exhibits convergence at a molecular level, and (4) all the mathematical of machinery of neutral theory works so well, THEN what magically prevents all these things from being true at the species level?
This all cannot be true for cancer, but false for evolution. That is the real inconvenience here.
The “real inconvenience”? After a snarky comment about ID from the same BioLogos reader, Swamidass elaborates:
Put another way, if many ID arguments in molecular biology were true, then cancer as we know it would be mathematically impossible, or regularly require the direct intervention of God to initiate and be sustained. I’m not sure which option is harder to believe.
To be clear, this does not in itself prove evolution is true. Rather, it casts serious doubt on the ID arguments from molecular biology (vis-a-vis Doug Axe and Kirk Durston, etc.).
So there you have it. The thread is long, but here are a few brief comments in response.
In Swamidass’s own paper in Nature Genetics, he develops a method to detect, identify, and classify mutations that occur commonly in known cancer-related genes. The goal seems to be to identify and classify cancer mutations that drive the cancer versus neutral mutations that don’t really have any effect. The paper tries to measure the “functional impact” that cancer mutations have. The ultimate goal, apparently, is to study individual patients and figure out what specific mutations caused their cancer.
The research is interesting. Let’s hope it bears fruit in healing or prevention. But what does it say about the power of mutations to generate something “novel”? The paper predicts:
It is biologically intuitive that there will be an interaction between mutation type and gene type in predicting drivers: we expect that truncations (frameshifts and premature stops) are less likely to be drivers than missense mutations [i.e., point mutations] when present in oncogenes, whereas the opposite is true in tumor-suppressor genes (TSGs), and silent mutations are unlikely to be drivers regardless of gene type.
What the paper says about tumor-suppressor genes makes sense. When tumor-suppressors are broken, as with a frameshift or premature stop, that totally disables their function, allowing tumors to grow. But those types of mutations are gene-killing mutations, and such mutations would not build anything new.
As to the prediction that oncogenes accumulate point mutations to become cancer drivers, that’s interesting, but from reading the paper one sees no discussion of any specific point mutations or what their specific phenotypic effect is in any specific genes. If Swamidass knows of examples that show “cancer regularly innovates with proteins of novel function,” his paper discusses none of those examples.
He does seem to indicate that cancer mutations can decrease protein specificity. But we already knew that mutations can do that from a 2015 paper by Douglas Axe and Ann Gauger. The problem, their paper found, is that mutations can’t build anything truly novel.
For details, see here:
Swamidass cites a Science paper from 1976, proposing the now-accepted hypothesis that cancer arises due to mutations in somatic cells (non-germ cells). This paper also proposed that cancer operates like Darwinian evolution. Except that, again, the mechanisms that cause cancer typically entail breaking features, not creating new ones
Swamidass gives a link to a 2012 Nature paper that reviews the cancer hypothesis proposed in the 1976 Science paper. This is a great review paper, expounding on a theory of how cancer is an evolutionary progression. However, it doesn’t spend much time discussing the specifics of those mutations and whether they are breaking molecular functions or building them.
Yet Swamidass says:
From this body of work, we can see the evolution of new functions (new information!), neutral theory, and the effectiveness of obtuse metrics like Ks/Ka ratios. It would hard to imagine rejecting evolution of species without somehow forgetting everything we have learned about the evolution of cancer.
In what way does cancer, a destructive disease, have anything to do with evolving new species? Cancer involves single cells, not whole organisms, and it doesn’t build new features, it tears down existing ones.
The argument from cancer doesn’t hold up. It doesn’t even make sense. “If many ID arguments in molecular biology were true, then cancer as we know it would be mathematically impossible,” writes Swamidass. Either that or it would “regularly require the direct intervention of God to initiate and be sustained.”
Not at all. “Things fall apart.” That is the natural way, which needs no evolutionary explanation. What ID and Darwinian theory seek to clarify is why things are built up — wonderful and functional things. Left to itself, the origami swan cited in Doug Axe’s book Undeniable will eventually crumple to powder. When its period on earth is completed, the living swan will decay too. How we get either in the first place is the question, and cancer surely isn’t the answer.