Detwiler: Questions Behe, Polyphen, and Ratchets


A protein doesn’t have enough mass to turn into a neutron star. The protein would have to have about 1.4 times the mass of the sun to collapse into a neutron star.

(John Detwiler) #78

" PolyPhen-2 is an automatic tool for prediction of possible impact of an amino acid substitution on the structure and function of a human protein. This prediction is based on a number of features comprising the sequence, phylogenetic and structural information characterizing the substitution."

" PolyPhen-2 checks if the amino acid replacement occurs at a site which is annotated as:


" PolyPhen-2 also checks if the substitution site is located in the region annotated as:


" PolyPhen-2 BLASTs query sequence against protein structure database ( PDB ) and by default retains all hits that meet the given criteria:

  • sequence identity threshold is set to 50%, since this value guarantees the conservation of basic structural characteristics"

" PolyPhen-2 uses DSSP ( D ictionary of S econdary S tructure in P roteins) database to get the following structural parameters for the mapped amino acid residues:

  • Secondary structure (according to the DSSP nomenclature)
  • Solvent accessible surface area (absolute value in Ų)
  • Phi-psi dihedral angles"

these are very general metrics that all relate to common structural elements in proteins. Just because your protein might have “improved function” does not mean that if a residue that use to make it’s core more hydrophic is changed to one that makes it less so all of a sudden doesn’t matter structurally.

I’m not neglecting negative selection. I don’t know why you think so. Polar bears are not somehow avoiding negative selection. Isn’t the whole point that these proteins were fixed precisely because of negative selection? What claim, precisely, do you think I’m making that is invalidated by the effects of negative selection?

(S. Joshua Swamidass) #79

Two things. I will be brief and to the point:

  1. Polyphen 2 does not predict stability. Quoting from the manual irrelevant text does not change this fact. It does not predict stability. Period.

  2. Negative selections weeds out mutations when they destabilizing too much. This prevents proteins from destabilizing entirely with time.

It does not appear you appreciate 1 or 2.

(John Detwiler) #80

So you agree that just because I didn’t vote for Hillary doesn’t mean I love Trump?

(John Detwiler) #81

It scores for progress toward known folds. Known folds have enough stability to fold. I think right now there are only just over 400 known folds. I don’t see how this wouldn’t be expected to correlate with stability even from just a sequence perspective.

Obviously, but your problem is not broken proteins that kill an organism, it’s the build up of many many almost broken proteins that still function well enough to keep an organism alive and well enough to reproduce. I didn’t post those links just to show that snps often break proteins. I think it’s a pretty safe bet that somewhere between being completely denatured and the original stability, there is an area of less stability that is still viable. But you knew this or you wouldn’t have brought up neutral evolution.

(S. Joshua Swamidass) #82

Why not just use a predictor designed to predict stability? There is just no point it treating an apple like its an orange, when an orange is easily available. This is pointless thing to resist. There is publicly available software to do this. Take a look at this:

Usually SNPs don’t break proteins. You know that right?

So the math of population genetics, and simulations too, shows that this isn’t a problem. When you are ready to learn about it we are here to teach you.


If one additional slightly broken protein prevents reproduction then that additional mutation will be strongly selected against. In your model, not all slightly deleterious mutations have the same effects since the most recent mutations will have a very strong deleterious effect.

In fact, this scenario is exactly what the evidence supports:

(John Detwiler) #84

I’m totally for checking it out. A couple nights ago I was on the site but I don’t have an academic email address for them to send the results. I also could not find in the paper where the precise index of the mutations were listed. They are illustrated graphically but in the downloadable methods section, I could not find an explicit description, nor would I expect one given the number of proteins and mutations. I don’t have time to do much more than copy and paste these things around atm.

actually I meant amino acid substitutions, which also don’t usually break a protein - only about 30% of the time.

I’m not against looking at it.

(S. Joshua Swamidass) #85

That is not true either. Did you follow what @Rumraket explained to you?

For humans, there are about 3 billion bases in the genome, and about 100 mutations each generation. Of these mutations, what percentage do you think are miss-sense mutations?

1 Like
(Mikkel R.) #86

Where have you published this conclusion?

I think Haldane’s dillema is still a real problem since it doesn’t seem to me that most larger metazoa have enough offspring to eliminate accumulating deleterious proteins let alone fix new more beneficial ones.


In fact, it demonstrably plays a role in our current fecundity.

What plays a role in our current fecundity? Homo sapiens is currently in an exponential population growth under relaxed selection due to agriculture and medical science. It would certainly be a mistake to extrapolate this pattern to the entirety of the biosphere.

I don’t even think recombination could get around this problem even without the effects of gene conversion which seem to operate many orders of magnitude faster.

So now you’ve told us what you think. Next you should do the work that shows your conclusions are correct and publish them.

1 Like

I couldn’t care less who you voted for. :sunglasses:

(Dave Carlson) #88

You can also download the software, install, and run it locally. Doesn’t seem like any email is required for this.
If you can’t/don’t want to do this, I can compile on my lab’s server and run whatever stability predictions you want. Just point me to the sequences you want to check.

(S. Joshua Swamidass) #89

@John_Detwiler, you can also use this one, from the lab in which I did my PhD. Believe it or not, computational modeling of biochemistry is in my wheelhouse.

1 Like
(John Detwiler) #90

I’m pretty sure that anything that prevents reproduction will be selected against.

I think you should say that mutations with strong deleterious effect are more likely to be recent. That is not saying the same thing.

It seems they did this by assuming the thing we are ultimately questioning if you read the age determination section. Also I don’t really think mutation rates inferred from common ancestry have fared well when measured directly.


Their age was determined by evidence, not simply assumed.

In this case, they are measuring the rate of base substitution, not necessarily the mutation rate. What they see is stronger conservation of sequence through time which supports the conclusion of stronger negative selection through time.

(John Detwiler) #92

Experimental evidence seems to show it’s true. Not sure which thing Rumraket said should convince me otherwise.

probably just under 3/4 from quickly looking at the table. Not sure how many are deletions and I think additions are rare. Why?


Hint: How many bases in the 3 billion base human genome code for an amino acid?

1 Like
(John Detwiler) #94

I’m not a writer or professional researcher. Sorry man.

Because of things like Behe’s rule. Because it seems that every day new disease causing mutations are arising, but we aren’t getting anywhere close to that number of new beneficial proteins found. Because of the fine tuned nature of protein protein interactions and the types of machines they produce. I think a lot of people imagine that building machines is easy if they never try. There’s a lot of other reasons. Right now I’m hoping to find out the locations of the mutations in these polar bear genes so I can check it in a more quantitative way.

I believe he was talking about negative selection. I don’t think it’s controversial that some people die because of mutations. Just because a population grows does not mean negative selection is not operating. I certainly think our genome can endure a good bit more degradation before advances in medicine and agriculture are unable to support more life than mutations prevent.

Why I already have a job. This is just a hobby.

(John Detwiler) #95

oh right, I was just thinking about coding regions. So about one percent of that figure then. Why?


It is important to the original question:

Those 100 mutations are spread across the entire genome. I think it would be more accurate to say that those 100 mutations are spread across the entire diploid genome which is 6 billion bases, and even that is a simplification since the father contributes many more mutations than the mother. Anyway . . .

Let’s just stick with 100 mutations across 3 billion bases to make it simple. If 1% of those bases actually code for an amino acid, what percentage of those mutations even have a chance of being missense mutations? I would say just 1 out of every 100 mutations affects a base in a coding region. We would also have to take 3rd base wobble into account, so less than 1 out of every 100 mutations will be a missense mutation. In fact, about 90% of mutations will occur in DNA with no DNA sequence specific activity so the vast majority of mutations are going to be neutral.

Does this sound right to you?