Optimization definitely applies to information theory.
1 Like
DNA is a molecule, not a machine. Arguments from analogy arenât going to work here.
3 Likes
Unless the environment the machine operates in changes. Then it may be rather easier to improve the machine. A Formula One car works great on a smooth paved racetrack. Try to run the same car in the off road Baja 1000 race and it wouldnât last a mile unless you modified the suspension, the tires, the gearing, raised the ground clearance, etc.
3 Likes
You still donât get it. A very slightly deleterious mutation doesnât break the metaphorical machine â that would be a substantially deleterious mutation. It might make the machine ever so slightly less efficient. If every possible mutation makes the machine less efficient, that means the machine was perfectly efficient to start with. No, you didnât say anything about the âmost functional possible stateâ, but GE requires that the genome be in its most functional possible state. Which means if you want to argue for GE, you have to explain why the genome was ever perfect.
5 Likes
Fun note, although this principle is intuitive, it has been supported by experiments testing fitness of gene variants under several different growth conditions. The conclusion - âacross all the environments and conditions that we tested, the wild type was the fittest alleleâ
5 Likes
To which of course Price and Sanford will reply âGod created all creatures with perfect genomes 6000 years agoâ. Then the ignoring of several hundred yearsâ worth of empirical scientific evidence completely falsifying the YEC rhetoric will continue.
2 Likes
I know they believe that, but GE is supposed to be an argument for creationism, not an argument that only makes sense if creationism is true.
5 Likes
Like all ID-Creationist ideas GE is actually an argument against evolution coupled with the false dichotomy âif evolution canât explain it then YEC must be trueâ. The concept of positive supporting evidence doesnât exist in ID-Creationist land.
1 Like
I really do get it. Do you?
Of course I understand that.
Yes, it likely will.
Nobody said anything about âevery possibleâ. But if you look at the scope of âall possible changesâ, then itâs beyond obvious that the vast majority of those possibilities will be on the spectrum of slight reductions of efficiency, not improvements. That is what is meant by:
Even the simplest of living organisms are highly complex. Mutationsâindiscriminate alterations of such complexityâare much more likely to be harmful than beneficial. - Gerrish, P. et al ., Genomic mutation rates that neutralize adaptive evolution and natural selection, J. R. Soc. Interface , 29 May 2013
The principle of âeasier to break than to improveâ still applies to small changes, just as it applies to large changes.
Thatâs simply not the case, and it was addressed directly in our joint article. GE will hold true even if the genome was never in a perfect state. There has never been a âperfect carâ, yet we still need to employ engineering and design principles to improve on cars. Itâs still easier to break a 1960s Ford by making arbitrary changes than it is to improve upon it.
Then given the rate GE is supposedly âdegradingâ the genome of every living species when in time (in years before present) did GE start taking effect? If you extrapolate back in time do you asymptotically approach a âperfectâ genome?
1 Like
Itâs beyond obvious you have no idea what youâre talking about. The vast majority of mutation are effectively neutral WRT fitness. That deleterious mutations greatly outnumber beneficial ones is only true in cases where a population is already very close to a local fitness maximum. But when the environment changes itâs not hard at all for mutations to find fitness improvements. Then natural selection does the rest.
2 Likes
Yes itâs almost like natural environments are generally more complex and varied than synthetic lab environments, and so the fittest wild-type allele is selected across a large range of fluctuating conditions over eons of time. /sarcasm
4 Likes
Better yet, imagine looking at single-locus protein-coding DFE and claiming it represents all loci. Imagine someone doing the experiment correctly through whole-genome interrogation and finding that environmental changes confer fitness advantage over the ancestor. Oh yeahâDillon et. al. did that in 2016.
Background: Spontaneous mutations in these bacteria are much more likely to produce deleterious mutations than humans and yet, the majority of mutations acquired in the experiment did not alter fitness. In the M9MM environment, 4 mutation carriers even had greater fitness than the ancestral genome. This means that effects of the mutations are dependent on the environment i.e.ânatural selection. Here are several quotes from that paper:
âSpecifically, MA experiments limit the efficiency of natural selection by passaging replicate lineages through repeated single-cell bottlenecks.â
âHere, we measured the relative fitness of 43 fully sequenced MA lineages derived from Burkholderia cenocepacia HI2424 in three laboratory environments after they had been evolved in the near absence of natural selection for 5554 generations . Following the MA experiment, each lineage harbored a total mutational load of 2â14 spontaneous mutations , including base substitution mutations (bpsms), insertion-deletion mutations (indels), and whole-plasmid deletions.â
âLastly, the genome of B. cenocepacia is composed of 6,787,380 bp (88.12%) coding DNA and 915,460 bp (11.88%) noncoding DNA . Although both bpsms and indels were observed more frequently than expected in noncoding DNA (bpsms: Ï2 = 2.19, d.f. = 1, P = 0.14; indels: Ï2 = 45.816, d.f. = 1, P < 0.0001).â
âIn combination, these results suggest that the fitness effects of a majority of spontaneous mutations were near neutral, or at least undetectable , with plate-based laboratory fitness assays. Given the average selection coefficient of each line and the number of mutations that it harbors, we can estimate that the average fitness effect ( s ) of a single mutation was â0.0040 ± 0.0052 (SD) in TSOY, â0.0031 ± 0.0044 (SD) in M9MM+CAA, and â0.0017 ± 0.0043 (SD) in M9MM .â
âDespite acquiring multiple mutations, the fitness of a number of MA lineages did not differ significantly from the ancestral strain. Further, the number of spontaneous mutations in a line did not correlate with their absolute selection coefficients in any environment (Spearmanâs rank correlation; TSOY: d.f. = 41, S = 15742, rho = â0.1886, P = 0.2257; M9MM+CAA: d.f. = 41, S = 13190, rho = 0.0041, P = 0.9793; and M9MM: d.f. = 41, S = 16293, rho = â0.2303, P = 0.1374).â
âBecause the fitness of many lineages with multiple mutations did not significantly differ from the ancestor, and because mutation number and fitness were not correlated, this study suggests that most of the significant losses and gains in fitness were caused by rare, single mutations with large fitness effects. â
âHere, we estimate that s â 0 in all three environments, largely because the vast majority of mutations appear to have near neutral effects on fitness . These estimates are remarkably similar to estimates from studies of MA lines with fully characterized mutational load in Pseudomonas aeruginosa and S. cerevisiae (Lynch et al. 2008; Heilbron et al. 2014), but are lower than estimates derived from unsequenced MA lineages (Halligan and Keightley 2009; Trindade et al. 2010).â
Dillon, M. M. & Cooper, V. S. The fitness effects of spontaneous mutations nearly unseen by selection in a bacterium with multiple chromosomes. Genetics 204 , 1225â1238 (2016)
3 Likes
Evidence that shows GE didnât occur is 100% on topic. That includes the huge amount of evidence life has been on the planet for over 3.7 billion years and hasnât gone extinct from GE yet. Like the 700,000 year old horse genome which has been sequenced. Evidence like that doesnât go away just because you canât explain it.
2 Likes
How does this refute GE exactly? I donât see it.
How does this refute GE exactly? I donât see it.
a) GE hypothesizes most mutations are bad, this is not observed experimentallyâeven when we control for natural selection, break the organismâs DNA repair capacity to generate more mutations than normal, use organisms with high coding and low non-coding genomes, and observe the population for several thousand generations
b) GE hypothesizes that fitness should decline, yet we see examples of fitness actually increasing
c) GE hypothesizes the organism will go extinct, yet there are no indications the organism goes extinct even after 5,000+ generations
d) GE hypothesizes that deleterious mutations are extremely common, yet the experiment demonstrates they are rare
e) GE hypothesizes fixed mutational effects on fitness, yet the experiment demonstrates that these effects differ depending on the environment
The predictions of GE are directly refuted via experimental conditions using bacteriaâwhich is why GE proponents make up special rules for microbes.
4 Likes
I believe you may be wrong about c) d) and e).
d) doesnât even seem to be right because the fitness effects look fairly balanced to me. How exactly does that prove that deleterious mutations are rare?
There can be special ârulesâ for microbes because nothing in the model requires their evolution to be the same as for other larger organisms with longer generation times. Itâs obviously not. It seems to strengthen their case that the organism most likely to beat GE just stays mostly neutral in terms of fitness when natural selection is not acting on it and it canât generate more mutations than normal. Itâs also possible that if it could generate more mutations, natural selection would act on it and it would remain mostly neutral.
But itâs late and my brain is dead, so Iâm not sure if Iâm making any sense, tbh. 
What about c) d) and e) do you disagree with?
Does GE hypothesize that an organism will go extinct due to accumulating mutations?
c) Did the mutations cause the bacteria to go extinct? No.
Does GE hypothesize that deleterious mutations are extremely common?
d) Did the experiment demonstrate an overwhelming majority of deleterious mutations? Noâthey found the opposite.
Does GE assume that fitness effects are static and unchanging regardless of the environment?
e) Did the experiment find differential fitness of the same mutations due to different environments? Yes.
The same mechanism (effective population size) that allows natural selection to prune deleterious mutations also works in humans. So a) there is no evidence that mutations are predominantly deleterious and b) natural selection works
That doesnât make sense. The exact mechanisms which allow bacteria to evade GE are also available and working in humans. The fact that we experimentally reduce natural selection and cause more mutations in the bacteria is strong evidence against the predictions of GE.
It might help to lay out the exact predictions of GE and what you would expect to find in an organism if GE were true.
GE requires natural selection to not actâthatâs the only way for those mutations to accumulate in the genome and cause extinction.
3 Likes
Sigh. Okay, great. Youâre not claiming it was in the best possible state, just that it was so close to perfect that the vast majority of possible changes were for the worse. Now that youâve made that pointless distinction, tell us why you think it was ever in that state. How did it get there? Whatâs the evidence that any organism was ever in such a state?
1 Like
Itâs not about whether it was ever in a best possible state. Itâs not about being âso closeâ to that state. Itâs about the fact that, in principle, itâs always easier to break a machine than it is to improve on it. A monkey can break a model locomotive. But can a monkey improve one? This has nothing to do with whether itâs a great machine. This holds true for any functional machine. Itâs a basic principle of engineering.
It seems like this is all going to boil down to the DFE. Will you grant that, if the DFE for effectively neutrals looks even remotely similar to the DFE for larger-sized mutations, then evolution as a theory needs to be thrown out?
Will you also grant along with Dr Dan Graur that, âIf ENCODE is right [i.e. if thereâs not much junk in the genome], then evolution is wrongâ?