Dennis Hackethal’s Blog
My blog about philosophy, coding, and anything else that interests me.
Why Do Humans Have Fewer Genes than Flies?
Genomics PhD Ruxandra Teslo asks:
[H]umans have only about 30,000-35,000 genes, two times less than a fly. Yet we are clearly more complex than a fly. How [is] that possible?
Creativity. That is, the ability to create new knowledge.
Put simply, organisms are computers that can move; their genes are the software. (That’s not a metaphor – it’s literally true.) In any non-creative organism, the onus is on genes to provide ~100% of the knowledge the organism requires to survive. But a creative organism, by definition, can come up with new knowledge during its lifetime.
Since a fly cannot create new knowledge ‘on the fly’ (pun intended), it needs ~all the knowledge it requires – flight, walking, etc – preinstalled in its genes. Likewise, a baby giraffe walks only a few hours after birth. Ducklings can walk and swim shortly after birth.1
To put it in the context of my Neo-Darwinian approach to the mind: one of our ancestors had a genetic mutation causing ideas to start replicating imperfectly inside his mind. This caused more evolution of knowledge to occur during this ancestor’s lifetime and took some of the onus off his genes. If an organism can create some knowledge itself then biological evolution doesn’t have to create that same knowledge anymore. Many species come out the womb ready to walk, but humans learn to walk themselves. So humans don’t need genes that would otherwise enable them to walk automatically.
Consider also that unfavorable genetic mutations are more numerous than favorable ones. Since creativity can make up for many unfavorable mutations, there was unprecedented selection pressure favoring creativity at roughly the rate at which unfavorable mutations occur. Creativity is a big deal from an evolutionary perspective. (Well, from ~any perspective, but especially from an evolutionary one.)
On top of that, using their creativity, humans create additional knowledge that genes wouldn’t have given them to begin with. Genes don’t code for nuclear reactors and spaceships, say. When a human comes up with a new idea, his knowledge has increased, but his genome has stayed the same.
Specifically, it’s the property physicist David Deutsch calls reach that is responsible for this big difference between the human genome and the genomes of all other species. The human genome codes for creativity, and creativity has more than enough reach to cover the knowledge of any other gene the genome could have contained (and even of all those it presumably could not have contained).
A multiplication table takes up more space on a piece of paper than a simple multiplication algorithm. Yet the algorithm gives you more results: it has greater reach. Comparing the number of genes is like comparing the number of lines of code: just because a program is longer doesn’t mean it does more. You need to compare reach instead.
Humans are the only creative animal. The human genome installs the basics, most importantly the creative algorithm, and then humans figure out the rest (and then some) during their lifetime.
So it’s not strange that humans have fewer genes than a fly. It’d be weird if they had more.
Teslo also asks:
Another perplexing finding was that most of the human genome was composed of so-called “junk DNA” — that is, DNA that did not code for any protein. The central dogma of Biology says that DNA codes for RNA, which in turn codes proteins, the molecules that carry out most of our cell’s functions. So what was all the DNA that was not coding for any protein doing there?
Well, maybe the central dogma is wrong. But let’s assume for a moment it’s correct.
Creativity explains this, too. Let’s take the example of walking again. Some of our creative ancestors presumably knew how to walk from birth: some of their genes pre-installed this knowledge in them. Then an erroneous mutation broke those genes, and with them, the inborn ability to walk. In any non-creative species, such a mutation would almost certainly lead to death, but in a creative species, it doesn’t have to: again, a creative organism can simply learn to walk. So error correction shifted from biological evolution to the organism itself. What had now become “junk DNA” could remain junk, and even grow and devolve further, without posing any serious risk to the remaining intact genes’ ability to spread. This process has been going on for hundreds of thousands of years, so our DNA has had plenty of time to accumulate junk.
In other words, the junk in our DNA is evidence of the fact that, for humans, error correction has largely shifted from biological evolution to the evolution of ideas inside human minds. A key Popperian insight (after philosopher Karl Popper) is that, when you want to know which of two processes creates more knowledge, you compare their error-correction rates.
So the answer to both questions is essentially the same.
Teslo raises these two questions in the wider context of figuring out why medicine and biology aren’t progressing faster. As you can see above, epistemology can solve biological questions. Maybe we need more epistemology.
-
Some animals require a certain level of physical maturing first, eg a fly larva cannot fly yet. But when an animal can do something ~immediately after birth, it means the knowledge for doing it is definitely inborn. ↩
Some animals display more flexibility by having so-called ‘learning’ algorithms driven by rewards and punishment, causing them to exhibit new behaviors not explicitly pre-installed. But those algorithms are different from creativity. New behaviors are not by themselves evidence of creativity.
Then there are animals which have fairly sophisticated imitation algorithms. For example, I understand kittens are not born with the ability to groom themselves and instead imitate their parents. But it’s still the imitation algorithm that’s inborn and ultimately responsible for the behavior. (It’s also worth noting that those imitation algorithms are often fairly indiscriminate and copy useless behavior – for example, I’ve seen videos of kittens that try to ‘bark’ because they were raised with dogs.)
One way to verify that some behavior is inborn and not ‘learned’ is by ‘raising’ an animal in isolation, ie without exposure to any members of the same species. For example, raise a fly larva by itself so it never sees any other flies. If it still flies once it has physically matured, then you know the behavior must be inborn and cannot be due to imitation algorithms, say. You can also rule out other types of ‘learning’ like play by physically restraining the larva. I suspect that a larva raised in isolation and restrained until maturity would fly immediately upon being released.
References
This post makes 2 references to:
- Post ‘Animal-Sentience FAQ’
- Post ‘The Neo-Darwinian Theory of the Mind’
There are 2 references to this post in:
What people are saying
Loved this article. Just a few thoughts.
Genes wouldn't be favored at roughly the rate of deleterious mutations because most genes don't code for learnable knowledge. For example, you can't learn to control the metabolic processes of your liver. In that sense, the selection pressure would be lower than the rate of deleterious mutations. On the other hand, the knowledge that can be learned creatively would also be able to be improved beyond what genetics can do. I'd imagine that the second of these forces far outweighs the first, so that the resulting pressure on creativity was much greater.
The 20,000 number is genes in the sense of genetics, that is - DNA that codes for proteins. Gene, in the neo-Darwininan sense, means a unit of natural selection, a part of a DNA strand that lasts enough generations for natural selection to exert selective pressure on it. Whether of not it actually codes for a protein is irrelevant. If strands of DNA that code for two proteins get brought together by linkage and become completely dependent on each other to replicate, those two strands become one gene in the evolutionary sense. There also must be code for what to do with the proteins. It's possible some of this code resides in the "junk DNA". To draw an analogy to computer programs - there can't be just data structures, there have to be actual algorithms. The "genetic code" is code for the data structures. We don't yet understand the code for the algorithms.
Reply
Thanks.
I think you mean creativity (or, more precisely, the genes coding for it) wouldn’t be favored and so on.
It’s true that most genes (in humans at least) don’t code for learnable knowledge, but creativity can make up for some physical shortcomings, too: if your genes give you a faulty leg, say, you can use your creativity to make a cane.
Having said that, you make a fair point. I’ve revised the article to be more specific about which mutations creativity can make up for. (I think one could control the metabolic processes of one’s liver by creating and taking the right medicine, but early humans obviously didn’t have that knowledge.)
Skipping some, you write:
Teslo spoke of “30,000-35,000 genes”, not 20,000.
Regarding the rest of your comment, epistemology also predicts that humans have more junk in their genes (in the neo-Darwinian sense) than any other species. And I could see missing or faulty protein synthesis leading to behavioral errors which creativity can then make up for. But I’ve edited the post to reflect the distinction you mention.
Reply