While on the interwebs I came across someone asking how costly junk DNA is. In other words, how much of a cell’s energy budget is taken up by replicating DNA that has no function. This caused me to do a bit of poking around. My initial reaction is that the cell makes way more RNA than DNA during its lifetime, and that seems to be the case. Eukaryotic genomes also require histones, so you have to account for the cost of maintaining histones. I did learn that histones have a relatively long half life (>100 days). Protein turnover differs greatly across proteins, as one would expect. I was surprised to see that muscle proteins like myosin have a half life measured in days.
To my eyes, the cost of junk DNA is pretty small compared to all of the other things a cell does. However, I finally came across a paper (thanks to Sandwalk) that sank its teeth into this question.
They were looking at the cost of a gene, so the full cost of replicating, transcribing, and translating a stretch of DNA. This can be easily applied to junk DNA by removing nearly all of the costs for transcription and translation. They found that the cost of protein translation is by far the highest cost, so not too surprising.
What did surprise me is that population genetics is probably the strongest driver of genome size. In my own mind, I had speculated that smaller genomes may be selected for on the basis of needing short generation times. This might explain why bacteria have small genomes, but why would algae have much larger genomes? Also, cell size plays a big role in that the cost of DNA replication is much smaller in larger cells that require more proteins and other structural features.
It’s not a big enough problem for it to be selected against in eukaryotes, if we are being overly accurate. Spurious transcription is very low cost just on the face of it given the low levels of expression involved. According to the numbers I have seen, the amount of RNA in a cell at any one given time is an order of magnitude higher than the amount of DNA. Of that RNA, only 1-3% is mRNA. The vast majority is ribosomal and transfer RNA. From what I have seen, spuriously transcribed RNA from junk DNA doesn’t even approach the expression seen in the vast majority of functional mRNA’s which themselves make up a tiny fraction of the RNA pool.
However, the cost of translation could have an impact on duplication of functional elements.
Neither is a lot of translation, for that matter. Evolution is terrible at turning off things where they aren’t needed. There’s no reason for an Intelligent Designer not to do so, though.
Yes it’s clear the cost of translation can in principle make a significantly bigger impact on overall cell energy budget than the mere cost of the replication of the extra DNA that encodes a duplicate gene.
In a way it’s too bad as I rather liked Nick Lanes’ idea about how mitochondria permitted expansion of the eukaryote genome over the comparatively lean prokaryote genome. But in the end, the trends are similar enough. Larger cells can afford to be larger and they can afford to have a lot of junk DNA.
But it’s still not clear to me why prokaryotes on the whole had not very much broken the large size barrier even though they had about 2 billion years before eukaryotes came on the scene.
Yes, I was aware of the basics that there are these super big bacteria. Then there is some eukaryotes that are very small, and they have small genomes to go with it.
What I understand from the argument of genetic load is that when we look at the functional regions in the genome, we find that most of the mutations that occur in these regions are harmful. If all the DNA were like these regions, there would be a quantity of mutations that humans could not tolerate. Therefore, most of the genome must be junk. My question now is whether everyone has mutations in these functional regions. If they do, then none of us will be able to survive because these mutations will not be inherited, and therefore we will become extinct. If someone says that natural selection will get rid of them, then we will all die because we all have harmful mutations, and this will lead to genetic decay. The only remaining option is that most of the mutations that occur in functional regions are not harmful, and only a very small percentage of them are harmful, which contradicts all scientific literature and argument of genetic load?" Is there a third option?
Suppose we have a population of organisms who all possess identical copies of a functional gene. This gene is highly intolerant of any mutations, such that any mutation whatsoever will be instantly lethal to the organism.
Now, what happens if one organism is born with a mutation to this gene? Answer: It instantly dies.
What happens to the mutated form of the gene? With the death of the organism, the mutated gene is instantly removed from the gene pool of the population. It cannot be passed on to any other organisms,
IOW, before the birth of the organism with this mutation, 100% if the members of the population have an identical, functional form of the gene. And after the birth (and subsequent death) of the mutant? Still 100%, because dead organisms are not part of the breeding population, for reasons that should be obvious.
Much like the cost of junk DNA, the cost of deleterious mutations is also a matter addressed by population genetics. If a genome has a high percentage of functional DNA then you also have to have a lot of offspring in each generation so that selection can filter out the deleterious mutations.
If 10% of the human genome is functional then the observed fertility rate in humans is enough for selection to remove the deleterious mutations, according to Graur’s calculations. Graur also estimates a 25% functional genome as an upper limit based on mutation rates and human fertility.
Let’s try some numbers. Every person has about 100 new mutations. If 10% of the genome is functional, then every person has around 10 mutations in functional regions. If every mutation in a functional region is deleterious, then every person has around 10 deleterious mutations. But that doesn’t mean the mutations are lethal. They could produce only a slight disadvantage. If there were no way to remove them, then over many generations there would indeed be genetic decay. But that’s where selection comes in, and it can remove a large number of harmful mutations at one time. If the entire genome were functional, that wouldn’t be enough, but apparently it can take care of mutations in that functional 10%. Also, it isn’t true that all mutations in functional regions are harmful. Based on the degree of conservation (a sign of selection operating), we might estimate that only at most half of mutations in the typical protein-coding exon are bad for you.
Don’t forget that an estimated 40% of human conceptions fail. They fail within a few days or weeks of fertilisation, often before the pregnancy is spotted, and that many/most of those are likely to be due to genetic issues. There is an awful lot of selection in humans that goes largely unnoticed.
From what I can see, both selection against weakly deleterious junk DNA and against much more strongly deleterious mutations in functional DNA shines a light fundamental mechanisms that govern both: population size and mutation rates. In bacteria there can be selection against relatively weak deleterious additions of junk DNA because of large population numbers. At the same time, mutation rates in bacteria are orders of magnitude lower than that found in humans (on a per nucleotide basis) which may be the result of selection for lower mutation rates in genomes with a much higher percentage of functional DNA.
Actually, it’s a function of the difference in what “per generation” means. A bacterial generation is one replication, while a human generation (especially in the male line) has orders of magnitude more. The fidelity of replication is about the same per replication.
This is really what I’m really thinking about, but I have a small problem, which is how natural selection will remove these mutations. All humans, in one way or another, will have harmful mutations. For example, I will have harmful mutations that I will pass on to my children. My children, in turn, will pass them on to their children. You also have harmful mutations. You will pass it on to your children, and every human being will have mutations and will pass them on, so how will natural selection get rid of them?
Well, drift will eliminate most of them. You will, on average, pass on only half of your mutations, and your children will pass on only half of that half, and so on. Some harmful mutations are eliminated by selection on gametes, others by selection on zygotes and embryos. Some are eliminated because people who have them will have fewer children than those who don’t. Also, consider that such mutations have been happening for billions of years, and we’re still around. It can’t be that big a problem.
Think of it in terms of more people being born than can survive. And that deaths in the population occur randomly(random sampling) but with the relative probability of survival being proportional to the fitness advantage of the alleles they carry.
In the long term the deleterious mutation will be outcompeted as the proportion of the population consisting of carriers of the fitter allele will get bigger and bigger while the proportion carrying the less fit allele will get smaller, because when it comes time to draw people for “random deaths”, a carrier of the less fit allele has a higher probability of being “sampled” than a carrier of the A allele.