Register | Sign In


Understanding through Discussion


EvC Forum active members: 63 (9161 total)
4 online now:
Newest Member: popoi
Post Volume: Total: 915,585 Year: 2,842/9,624 Month: 687/1,588 Week: 93/229 Day: 4/61 Hour: 0/0


Thread  Details

Email This Thread
Newer Topic | Older Topic
  
Author Topic:   Do you really understand the mathematics of evolution?
Kleinman
Member (Idle past 325 days)
Posts: 2142
From: United States
Joined: 10-06-2016


Message 1 of 239 (876413)
05-16-2020 1:37 PM


If you think you understand the mathematics of evolution, please explain the mathematics of the Kishony Mega-Plate experiment and the Lenski Long Term Evolutionary Experiment and what is the significant mathematical difference between the two experiments.

Replies to this message:
 Message 2 by Admin, posted 05-17-2020 8:59 AM Kleinman has replied
 Message 25 by Taq, posted 05-21-2020 5:16 PM Kleinman has replied

  
Kleinman
Member (Idle past 325 days)
Posts: 2142
From: United States
Joined: 10-06-2016


Message 3 of 239 (876415)
05-18-2020 2:49 AM
Reply to: Message 2 by Admin
05-17-2020 8:59 AM


Sorry, I thought the bloggers on this forum would be familiar with these two experiments. You can learn about the Kishony Mega-Plate experiment here:
Spatiotemporal microbial evolution on antibiotic landscapes - PMC
and you can watch a video of the experiment here:
https://www.youtube.com/watch?v=Irnc6w_Gsas
Essentially, what these experimenters do is make a large petri dish. And on this petri dish, they put different levels of an antibiotic in bands on the dish. On the left and right edges of the petri dish, no antibiotic. As you move toward the middle of the petri dish, the antibiotic concentration is increased in each band. Bacteria are then inoculated on the petri dish which initially grows only in the drug-free region. But as colonies grow, variants appear which can grow in the next higher drug concentration region. In this experiment, the antibiotic is the selection pressure.
The Lenski Long Term evolution experiment is performed in a different manner. You can read about how this experiment performed here:
E. coli long-term evolution experiment - Wikipedia
There many other publications based on this experiment but this link gives the essential principles. Essentially, what this experiment consists of are vials which contain 10ml of glucose growth solution for bacteria. Each day 5e6 bacteria are introduced into the solution which has sufficient nutrients to support the growth of about 5e8 bacteria (about 6-7 doublings (generations) per day). By the end of the day, the glucose is used up. 1% of the bacteria are then taken from this solution and introduced into a fresh 10ml solution for the next day's growth. Every 500 generations, the bacteria are saved, their genetic sequence determined, mutations identified. In this experiment, the selection pressure is starvation (nutrient limitation). This experiment has been carried out continuously since 1988 until it was paused recently because of the Covid episode.
If you or any members of this forum know how to correctly describe the physics and mathematics of either of these experiments or know of links that do this, please provide that information.

This message is a reply to:
 Message 2 by Admin, posted 05-17-2020 8:59 AM Admin has replied

Replies to this message:
 Message 4 by Admin, posted 05-18-2020 8:03 PM Kleinman has replied
 Message 7 by AZPaul3, posted 05-19-2020 12:20 PM Kleinman has replied

  
Kleinman
Member (Idle past 325 days)
Posts: 2142
From: United States
Joined: 10-06-2016


Message 5 of 239 (876417)
05-19-2020 7:52 AM
Reply to: Message 4 by Admin
05-18-2020 8:03 PM


Both.

This message is a reply to:
 Message 4 by Admin, posted 05-18-2020 8:03 PM Admin has not replied

  
Kleinman
Member (Idle past 325 days)
Posts: 2142
From: United States
Joined: 10-06-2016


Message 8 of 239 (876428)
05-19-2020 4:12 PM
Reply to: Message 7 by AZPaul3
05-19-2020 12:20 PM


If this PDF is your work, and I suspect it is, then you already have your answers since you wrote the paper that answered them.
It is. And I've also written a paper on the mathematics of the Lenski paper. You can find that paper here:
Just a moment...
And yes, I'm pretty sure that I understand the physics and mathematics of DNA evolution.
Why are you here?
I suspect there is a deeper point you want to make with your OP.
What would that be?
I'm just finishing another paper. I'm using a different approach to solve the mathematics of DNA evolution. Rather than doing the math as a simple probability problem using the "at least one rule", I'm using a Markov Chain. If you are not familiar with that math, you can read about it here: Models of DNA evolution - Wikipedia . I'm in the process of writing the discussion section and since I'm not in academia, I'm looking to discuss the subject before I submit it to the peer reviewers.
So, do you really understand the mathematics of evolution, in particular, DNA evolution?

This message is a reply to:
 Message 7 by AZPaul3, posted 05-19-2020 12:20 PM AZPaul3 has replied

Replies to this message:
 Message 9 by AZPaul3, posted 05-19-2020 5:09 PM Kleinman has not replied
 Message 10 by Kleinman, posted 05-19-2020 7:59 PM Kleinman has not replied
 Message 46 by mike the wiz, posted 05-28-2020 6:09 AM Kleinman has replied

  
Kleinman
Member (Idle past 325 days)
Posts: 2142
From: United States
Joined: 10-06-2016


Message 10 of 239 (876434)
05-19-2020 7:59 PM
Reply to: Message 8 by Kleinman
05-19-2020 4:12 PM


First, if you are looking for some pre-review help on the mathematical efficacy of this new analysis of yours then you are in the wrong forum. A math or a biology forum would seem to be more beneficial.
Second, if this isn't to be a rehash of the discussion in Explaining the pro-Evolution position from years ago then what differences from that discussion are you offering?
So you don't really understand the mathematics of evolution. Maybe someone else wants to try and explain the Kishony and Lenski experiments. What is the correct explanation for why it takes a billion replications for each evolutionary step in the Kishony experiment? Perhaps someone wants to try and explain why drug-resistant variants appear in Lenski's founder's population despite the fact that these bacteria were never exposed to antibiotics? Or perhaps you want to explain why competition slows evolutionary adaptation?

This message is a reply to:
 Message 8 by Kleinman, posted 05-19-2020 4:12 PM Kleinman has not replied

Replies to this message:
 Message 11 by nwr, posted 05-19-2020 8:53 PM Kleinman has not replied
 Message 12 by AZPaul3, posted 05-19-2020 9:58 PM Kleinman has not replied
 Message 26 by Taq, posted 05-21-2020 5:26 PM Kleinman has not replied

  
Kleinman
Member (Idle past 325 days)
Posts: 2142
From: United States
Joined: 10-06-2016


Message 14 of 239 (876454)
05-20-2020 10:53 AM


Responses nwr and AZPaul3
nwr writes:
I am puzzled by your repeated references to "mathematics of evolution."
Are you familiar with the work of Haldane, Kimura, and many other population geneticists?
Kleinman writes:
What is the correct explanation for why it takes a billion replications for each evolutionary step in the Kishony experiment?
nwr writes:
Under what definition of "correct"?
It takes as many replications as it takes. This was an empirical experiment to observe how organisms behave.
So why doesn't it take 10 or 100 replications for each evolutionary step?
Kleinman writes:
Perhaps someone wants to try and explain why drug-resistant variants appear in Lenski's founder's population despite the fact that these bacteria were never exposed to antibiotics?
nwr writes:
Why does that need explaining? Again, this was an empirical investigation to see how biological organisms behave.
I'm completely puzzled about what you are looking for. I do not see any point to the questions that you are asking.
Don't you find this surprising that drug-resistant bacteria would appear in bacterial populations that were never exposed to the antibiotics and there was resistance to many different antibiotics?
Kleinman writes:
What is the correct explanation for why it takes a billion replications for each evolutionary step in the Kishony experiment?
AZPaul3 writes:
Does it? Unless I missed something I don't see that in Kishony's paper.
Even if that were the case is that significant in some way?
You did see the part where they describe
Hence, as compensatory mutations often occur behind the front, they are spatially restricted from contributing to the ultimate evolutionary course of the population.
Source
You are correct, neither Kishony nor any members of the Kishony team have ever published a paper describing the mathematics of their experiment. Kishony only talks about the billion replications in his video which I linked to in Message 3.
However, are you claiming that it takes fewer than a billion replications for compensatory mutations to occur because these variants are appearing behind the front but their migration is inhibited? If so, how many fewer replications are needed?
AZPaul3 writes:
In other words because of the experimental setup the more resistant bugs often got blocked and had they not been the population would have entered the next gradient much more quickly.
When you say more quickly, do you mean fewer replications for each lineage on their particular evolutionary trajectory or just a shorter duration of the experiment if you reduce the migration time of the variants with the compensatory mutations?
Kleinman writes:
Perhaps someone wants to try and explain why drug-resistant variants appear in Lenski's founder's population despite the fact that these bacteria were never exposed to antibiotics?
AZPaul3 writes:
You don't understand random mutation? You don't understand evolution.
What makes you think exposure is necessary to develop a resistance capability in an individual or a small subset of a population? Do you not know why genetic diversity is so powerful an evolutionary tool? You don't understand evolution.
I never said exposure was required for drug-resistance to appear. It is simply a matter of population size and mutation rate which determines the appearance of these variants. So, why don't you tell us how large the population must be for these drug-resistant variants to appear?
Kleinman writes:
Or perhaps you want to explain why competition slows evolutionary adaptation?
AZPaul3 writes:
Since both of these experiments were set up in highly restrictive environments for highly restrictive purposes, trying to conclude this is a general feature within the global biosphere of evolution's reality is not just bogus it is dishonest.
You're trying to use your math manipulations (mathnipulations ) as a sword to strike at the windmill of evolution?
My advice is go learn something about the subject first.
You can't even do the mathematics of these experiments "set up in highly restrictive environments". Why should anyone believe that you can correctly explain the mathematics of evolution in unrestricted environments?
So, when are you going to design an experiment that works more quickly than either the Kishony or Lenski experiments? You should tell them to stop restricting their experiments.

Replies to this message:
 Message 15 by nwr, posted 05-20-2020 12:40 PM Kleinman has not replied
 Message 16 by AZPaul3, posted 05-20-2020 12:41 PM Kleinman has not replied
 Message 17 by NosyNed, posted 05-20-2020 1:22 PM Kleinman has not replied
 Message 23 by ringo, posted 05-21-2020 11:50 AM Kleinman has not replied

  
Kleinman
Member (Idle past 325 days)
Posts: 2142
From: United States
Joined: 10-06-2016


Message 18 of 239 (876475)
05-20-2020 5:29 PM


Responses nwr, AZPaul3, and NosyNed
Kleinman writes:
Are you familiar with the work of Haldane, Kimura, and many other population geneticists?
nwr writes:
Somewhat. The mathematics that they use is ordinary mathematical probability theory. Is that what you mean by "mathematics of evolution"?
The mathematics of evolution requires different mathematics depending on the evolutionary process that you are trying to describe. You can't use the mathematics of DNA evolution to describe competition and fixation but you can have both processes occurring simultaneously such as what is occurring with the Lenski experiment. And the mathematics of recombination is different than either DNA evolution or competition and fixation. Mendelian genetics is a simple example of a recombination model. DNA evolution is stochastic (it can be modeled as a random walk or done as a simple probability problem using the "at least one rule"). Competition and fixation, on the other hand, is deterministic (unless you are talking about fixation by drift).
Kleinman writes:
So why doesn't it take 10 or 100 replications for each evolutionary step?
nwr writes:
I puzzled as to why you would think that.
Evolution isn't a mechanical system to grind out a sequence of steps leading to a particular result. Rather, evolution is a system of keeping the population well adapted to a changing environment.
But that's exactly what the Kishony experiment is demonstrating. His lineages are grinding out replications until some lucky member gets the correct mutation that enables that variant to grow in the next higher concentration drug region. Then that new variant must grind out more replications until some lucky descendent gets the next beneficial mutation that enables that new variant to grow in an even higher drug concentration region. It's a random walk with lots of members taking random steps but only the lucky few get to grow in the next higher drug concentration regions.
Kleinman writes:
Don't you find this surprising that drug-resistant bacteria would appear in bacterial populations that were never exposed to the antibiotics and there was resistance to many different antibiotics?
nwr writes:
No, that is not at all surprising.
Evolution is a system of trial and error, in order to see what works well. If you are using trial and error methods, then you are going to try out things that turn out to not work. So the bacteria population tries mutations that turn out to not be currently helpful. That's what we should expect.
If they could not produce occasional antibiotic resistant variants when not useful, then they also could not produce them when would be useful.
Good, so now learn how to do the mathematics of this process. Try to understand why it takes a billion replications for a lineage to accumulate each particular mutation that gives resistance to the antibiotic used in the Kishony experiment. And by the way, the same principle applies to the lineages in the Lenski experiment, however, they are not accumulating the mutations which give drug resistance (in fact, those mutations disappear from his populations), Lenski's lineages are accumulating mutations which give more efficient usage of the limited glucose available.
Kleinman writes:
So why doesn't it take 10 or 100 replications for each evolutionary step?
AZPaul3 writes:
What evolutionary step? The edge of the next gradient? Do you know how far that is? Do you know the size of the bacteria?
What makes you think resistance was not achieved in the first 10 generations in some subset of the population? What makes you think that resistance wasn't lost again in the next 10? What makes you think resistance could not have been developed and lost many times by many disparate subsets of the population in this long slog across the gradient?
You don't know how evolution works do you?
You are conflating geometry and genetics. The correct mutation which would allow a variant to grow in the next higher concentration region could occur in 10 or 100 replications, but it might take more than a billion replications as well. But try and recall what the meaning of a mutation rate is and that is a binomial probability problem (does the beneficial mutation occur or not). And if the mutation rate is 1e9, you will have on average only one occurrence of that particular mutation every billion replications. Since this is a binomial probability problem, check and see what the standard deviation and variance are for this problem and why the number of replications is much more likely to be close to a billion replications.
Kleinman writes:
Don't you find this surprising that drug-resistant bacteria would appear in bacterial populations that were never exposed to the antibiotics and there was resistance to many different antibiotics?
AZPaul3 writes:
If you understood how evolution worked you wouldn't find such an occurrence surprising either. Just because the natural selection side of the process isn't being tested yet doesn't mean the random mutation side doesn't still occur.
Thus is the power of accumulating genetic diversity in a large population.
I don't and didn't find this surprising but Lenski did. That's why he wrote a paper on this issue and he didn't explain why it happens. And if you consider the Kishony experiment when he used Ciprofloxacin, his population of a billion not only will have a high probability of having the first beneficial mutation for Ciprofloxacin resistance, he also has a high probability of a variant with a beneficial mutation for Trimethoprim (which won't help that variant in the Ciprofloxacin environment).
Kleinman writes:
You are correct, neither Kishony nor any members of the Kishony team have ever published a paper describing the mathematics of their experiment. Kishony only talks about the billion replications in his video which I linked to in Message 3.
AZPaul3 writes:
Your citing as their scientific data the informal talk of the group's PopSci outreach video? Academic quality noted.
Hey, you run with the data you have. Would you be more comfortable with the Lenski experiment data? In his experiment, he gets fixation for each beneficial mutation in a range of 200-1000 generations/mutation. He gets about 5e8 replications/day based on about 6 1/2 doublings/day (generations). So, how many replications for each beneficial mutation in the Lenski experiment?
Kleinman writes:
So, why don't you tell us how large the population must be for these drug-resistant variants to appear?
AZPaul3 writes:
From a minimum of 1 to an indeterminate maximum. And you don't understand why that is do you.
I can narrow down your estimate from 1 to infinity a little bit and you could as well if you understood how to do the mathematics of evolution.
Kleinman writes:
Don't you find this surprising that drug-resistant bacteria would appear in bacterial populations that were never exposed to the antibiotics and there was resistance to many different antibiotics?
NosyNed writes:
Why would that be at all surprising? That question of yours makes it sound like you expect a mutation to appear because it is needed. But that has nothing to do with it. I'm being silly but it almost makes it seem like you have no clue about evolution.
I've never said that. I know that mutations are random events. So, now that you understand why you don't need antibiotics for drug-resistant variants to appear, why should medical providers be discouraged from using antibiotics?

Replies to this message:
 Message 19 by jar, posted 05-20-2020 5:33 PM Kleinman has replied
 Message 20 by nwr, posted 05-20-2020 6:24 PM Kleinman has not replied
 Message 22 by AZPaul3, posted 05-20-2020 7:45 PM Kleinman has not replied

  
Kleinman
Member (Idle past 325 days)
Posts: 2142
From: United States
Joined: 10-06-2016


Message 21 of 239 (876479)
05-20-2020 6:34 PM
Reply to: Message 19 by jar
05-20-2020 5:33 PM


Time for jar to sober up
Kleinman writes:
I know that mutations are random events. So, now that you understand why you don't need antibiotics for drug-resistant variants to appear, why should medical providers be discouraged from using antibiotics?
jar writes:
That alone should be more than enough to simply close this thread down and adjourn for a pint.
Id10T
When you sober up, why don't you tell us what is the most common medical reason for hospital admission? Since you still probably won't see straight after you sober up, here's the data:
Most Common Diagnoses in Hospital Inpatient Stays - HCUP Fast Stats

This message is a reply to:
 Message 19 by jar, posted 05-20-2020 5:33 PM jar has not replied

  
Kleinman
Member (Idle past 325 days)
Posts: 2142
From: United States
Joined: 10-06-2016


Message 27 of 239 (876538)
05-21-2020 7:19 PM
Reply to: Message 25 by Taq
05-21-2020 5:16 PM


Something has come up and I'll busy for the next few days but I wanted to respond to Taq's message. I'll try to get to the others next week.
Kleinman writes:
If you think you understand the mathematics of evolution,
Taq writes:
Population genetics isn't one of my strong suits, but I am acquainted with it.
Is there something specific you wish to discuss? A lot of the math can be found here:
404
Yes there is a lot of math. So, how do you apply it to the Kishony and Lenski experiments?
Kleinman writes:
please explain the mathematics of the Kishony Mega-Plate experiment and the Lenski Long Term Evolutionary Experiment and what is the significant mathematical difference between the two experiments.
Taq writes:
At first glance, the major differences would be population dynamics. In the Lenski experiment you had a severe population bottleneck every 10 or so generations. This large fluctuation in effective population size would certainly have effects on genetic drift.
The bottleneck in the Lenski experiment is about every 6 1/2 generations. What that does is introduce a discontinuity in the mathematics which makes it very difficult to model this with a closed-form solution. You can see how I handled the discontinuity using a numerical solution here:
Just a moment...
But there is another major difference between the Kishony and Lenski experiments. The Kishony experiment has a much larger carrying capacity than the Lenski experiment. What this does is it forces his populations to compete for the limited resources and the more fit variant must drive the less fit variant to extinction in order to accumulate the necessary replications for improving fitness. This slows the DNA evolution when compared with the rate of evolution in the Kishony experiment. And note, the less fit variants in the Kishony experiment are still living on in the lower drug concentration regions, no fixation required for adaptation. The mathematics for DNA evolution is still the same for both, but you can see why competition slows the evolutionary process. The less fit variants are consuming resources that the more fit variant could use to replicate and will continue to do so until these variants are driven to extinction. By the way, I show in the link above how to predict what would happen if Lenski used 1ml or 100ml of the solution instead of the 10ml.
What do you think would happen if Lenski were to run his experiment at non-optimal temperature (thermal stress) along with the starvation stress? Or if Kishony were to try to run his experiment with 2 drugs instead of 1. We already know that the Kishony experiment won't work as designed unless the increase in drug concentration is limited so that a single mutation gives adaptation to the next higher drug concentration region. What would it take to make that experiment work?
Have a good weekend, see y'all next week.

This message is a reply to:
 Message 25 by Taq, posted 05-21-2020 5:16 PM Taq has replied

Replies to this message:
 Message 28 by Taq, posted 05-22-2020 11:10 AM Kleinman has not replied

  
Kleinman
Member (Idle past 325 days)
Posts: 2142
From: United States
Joined: 10-06-2016


Message 29 of 239 (876662)
05-25-2020 10:41 AM


Responses to Taq and several other members
AZPaul3's edited version of what Kleinman writes:
His lineages are grinding out replications until ...
AZPaul3 writes:
No "until". The genetic grind never stops and it certainly wasn't aimed at the next "step".
Try quoting my entire statement then perhaps you may understand what I'm saying. Here's my full quote.
Kleinman writes:
His lineages are grinding out replications until some lucky member gets the correct mutation that enables that variant to grow in the next higher concentration drug region.
And sure the grind continues (unless extinction occurs which is the most common outcome for most lineages). Too bad you don't how to do the mathematics of this grind, then you could explain the mathematics of the evolution of drug resistance.
Kleinman writes:
Don't you find this surprising that drug-resistant bacteria would appear in bacterial populations that were never exposed to the antibiotics and there was resistance to many different antibiotics?
Ringo writes:
I only claim to have the first inkling of a clue about evolution but I don't find that surprising. Drug-resistant bacteria evolve, whether there are drugs in the neighbourhood or not. Hint: mutations. When drugs do mosey by, the drug-resistant bacteria survive. No mystery.
It is not a mystery to me as well. So why should physicians be discouraged from using antibiotics when you don't need the antibiotics for resistant variants to appear? This is an especially important question when septicemia and pneumonia are the #1 and #4 causes for hospital admission which could be prevented with timely usage of oral antibiotics in the outpatient environment before these infections become severe.
Kleinman writes:
Perhaps someone wants to try and explain why drug-resistant variants appear in Lenski's founder's population despite the fact that these bacteria were never exposed to antibiotics?
Taq writes:
This was explained over 60 years ago. The answer is random mutations that occur in the background.
Esther and Joshua Lederberg worked out the concept of spontaneous mutations and how it results in antibiotic resistance, and you can read their 1952 paper here:
REPLICA PLATING AND INDIRECT SELECTION OF BACTERIAL MUTANTS - PMC
Luria and Delbruck worked out all of the math for the process in this paper from 1943:
https://www.genetics.org/content/28/6/491
OK, now use this math to explain the behavior of the Kishony and Lenski experiments. And then, using those equations, predict the behavior of the Lenski experiment if he uses 1ml or 100ml of solution instead of 10ml and predict the behavior the Kishony experiment if he were to try to use two drugs instead of one.
Taq writes:
Interestingly, the math of evolution and the math of slot machines are somewhat related:
Actually, the mathematics of evolution is more like the math of coin tossing. The mathematics of evolution consists of nested binomial probability problems where each binomial probability problem is joined to the other by the multiplication rule of probabilities.

Replies to this message:
 Message 30 by nwr, posted 05-25-2020 9:58 PM Kleinman has replied

  
Kleinman
Member (Idle past 325 days)
Posts: 2142
From: United States
Joined: 10-06-2016


Message 31 of 239 (876708)
05-26-2020 9:31 AM
Reply to: Message 30 by nwr
05-25-2020 9:58 PM


Trying to give nwr more than a vague understanding of DNA evolution
Kleinman writes:
Try quoting my entire statement then perhaps you may understand what I'm saying. Here's my full quote.
Kleinman writes:
His lineages are grinding out replications until some lucky member gets the correct mutation that enables that variant to grow in the next higher concentration drug region.
Kleinman writes:
And sure the grind continues (unless extinction occurs which is the most common outcome for most lineages).
nwr writes:
I guess it is hopeless. You are continuing to make the same mistake that we often see being made by creationists. And you are damned sure that you are right, even though you are wrong.
When you said in response to my question previously:
Kleinman writes:
Are you familiar with the work of Haldane, Kimura, and many other population geneticists?
nwr writes:
Somewhat. The mathematics that they use is ordinary mathematical probability theory. Is that what you mean by "mathematics of evolution"?
You are not as familiar with "ordinary" mathematical probability theory as you think you are.
Kleinman writes:
Actually, the mathematics of evolution is more like the math of coin tossing. The mathematics of evolution consists of nested binomial probability problems where each binomial probability problem is joined to the other by the multiplication rule of probabilities.
nwr writes:
No, that's wrong.
It is you who is wrong nwr. DNA evolution is a binomial probability problem. The random trial is a replication and the two possible outcomes are, does a beneficial mutation occur or does a beneficial mutation does not occur. That is analogous to a coin toss problem where the random trial is the toss of the coin and the two possible outcomes does a head occur or does a head not occur. The major difference between the two examples is that the coin tossing problem is symmetric, ie, the possible outcomes have equal probabilities of 0.5. The DNA evolution binomial probability example is highly asymmetric, that is, the probability of a beneficial mutation occurring in a single replication is the (beneficial) mutation rate and the probability of a beneficial mutation not occurring in a single replication is 1 minus the (beneficial) mutation rate.
nwr writes:
That's about what you would get if the environment were completely static. But the environment is changing all the time. So the probabilities are changing all the time.
It doesn't matter whether the environment is static or not. The DNA evolutionary process still consists of nested binomial probability problems. The Kishony and Lenski experiments are as close to static as you can get. They hold all their variables including the selection conditions as constant as possible and it still takes billions of replications for each step improvement in fitness for their variants. If they were to change their environments at any time during the performance of the experiment, that would mean they would be changing the evolutionary trajectories for increasing fitness making those evolutionary trajectories more complex. That's why Kishony hasn't been able to get his experiment to work with two drugs (or if the step increase in drug concentration is too large) and Lenski's experiment would work even more slowly if he were to do it at a non-optimal temperature. The mathematical explanation for this is the multiplication rule of probabilities.
nwr writes:
The problem of drug resistant bacteria isn't that they show up. The problem is too much use of antibiotics will change the environment so that antibiotic-resistant bacteria become more highly favored. And then we will see many more of them and they become a problem.
You are half right here. If antibiotics are used incorrectly, what you will do is colonize your patients with drug-resistant variants. The way you do this is by using single-drug therapy on someone who is immune-compromised. What happens is you kill off the drug-sensitive variants but because the person's immune system doesn't remove the remaining drug-resistant variants, not only will you have treatment failure but the person will now be colonized by bacteria resistant to that drug. And as that colony grows, there is an increasing probability that more variants will appear resistant to higher concentrations of the drugs (just as Kishony's variants are moving to the higher drug concentrations regions).
I don't know if you paid attention to my responses to jar. Withholding antibiotics from someone who needs them not only has health consequences but economic consequences as well. Septicemia and pneumonia are the number 1 and 4 medical reasons for hospitalization. Do a search on the hospitalization cost for the treatment of those diseases and this problem is getting worse. Many of these cases could be prevented with the timely and correct usage of antibiotics with much lower costs in the outpatient environment. It is the incorrect or in your case vague understanding of evolution that is contributing to this problem.
My suggestion to you is if you want to have a better understanding of this is to watch the Kahn Academy (or various other) lectures on probability theory on YouTube. Then read this link:
Page not found - Stats & Data Science Views
And when you understand that, read this paper:
Just a moment...
When you do that you might have a better idea of what I'm talking about.

This message is a reply to:
 Message 30 by nwr, posted 05-25-2020 9:58 PM nwr has replied

Replies to this message:
 Message 32 by Phat, posted 05-26-2020 4:47 PM Kleinman has replied
 Message 34 by Taq, posted 05-26-2020 5:35 PM Kleinman has replied
 Message 36 by nwr, posted 05-26-2020 9:23 PM Kleinman has not replied

  
Kleinman
Member (Idle past 325 days)
Posts: 2142
From: United States
Joined: 10-06-2016


Message 33 of 239 (876721)
05-26-2020 5:29 PM
Reply to: Message 32 by Phat
05-26-2020 4:47 PM


a distinct difference between chance & true randomness vs a set & measurable prob?
Phat writes:
all that i would add, not being a scientist or a creationist is that there is a distinct difference between chance and true randomness versus a set and measureable probability.
I think you are trying to draw a distinction using semantics when there is no mathematical difference. But, if you think I'm wrong, give us empirical examples of "chance and true randomness" and a "set and measurable probability" and show us what that distinct difference is.

This message is a reply to:
 Message 32 by Phat, posted 05-26-2020 4:47 PM Phat has not replied

  
Kleinman
Member (Idle past 325 days)
Posts: 2142
From: United States
Joined: 10-06-2016


Message 35 of 239 (876729)
05-26-2020 7:15 PM
Reply to: Message 34 by Taq
05-26-2020 5:35 PM


Re: Trying to give nwr more than a vague understanding of DNA evolution
Kleinman writes:
It is you who is wrong nwr. DNA evolution is a binomial probability problem. The random trial is a replication and the two possible outcomes are, does a beneficial mutation occur or does a beneficial mutation does not occur. That is analogous to a coin toss problem where the random trial is the toss of the coin and the two possible outcomes does a head occur or does a head not occur.
Taq writes:
No analogy is perfect, but I don't think a coin toss is a good one. First, the benefice of a mutation can be small or great, not simply heads or tails. Second, the same mutation can be neutral or detrimental in one genetic background and beneficial in another. A lottery might be a better comparison where someone can win 5 bucks or hundreds of millions, and the winning ticket from one drawing won't necessarily be a winner in the next.
That's correct that a mutation can be neutral or detrimental in one environment (what you call a genetic background) and beneficial in another. But that doesn't make any difference in the mathematics of the evolutionary trajectory. Think of it in the context of the Kishony experiment. When Kishony is using ciprofloxacin as the selection pressure. His variants on the ciprofloxacin evolutionary trajectory are accumulating the mutations necessary for adaptation by a billion replications per evolutionary step until one of the lucky members gets the next beneficial mutation to take a step on that evolutionary trajectory and grow in the next higher drug concentration region. In his population are other members getting mutations that are beneficial for other drugs (such as trimethoprim) but his petri dish is not large enough for those variants to accumulate the billion replications for the trimethoprim resistance evolutionary trajectory. If Kishony made an even larger petri dish without any antibiotic such that he could get much, much, larger populations, he would get variants resistant to all kinds of drugs, just as Lenski's founder's populations did. Every evolutionary trajectory consists of nested binomial probability problems joined by the multiplication rule. What determines the ability of a lineage to take a particular evolutionary trajectory is the ability of each variant at each evolutionary step to replicate sufficiently to have a reasonable probability to get a beneficial mutation. And when the probability of success is given by the mutation rate, it's going to be a lot of replications at each evolutionary step as demonstrated by the Kishony and Lenski experiments.
Kleinman writes:
It doesn't matter whether the environment is static or not.
Taq writes:
Umm, yes it does. Read up on fitness landscapes and niche specialization. As you near an optimum the number of beneficial changes is greatly reduced. A changing environment results in optimized species being less optimize which increases the number of potential beneficial changes.
Umm, no it doesn't. What changes could you make to the Kishony experiment to make his populations evolve more quickly? And if you want to make the Lenski experiment evolve more rapidly, increase the carrying capacity. That would reduce the competition allowing the adapting lineage to achieve more replications in the same time interval. But if any changes to the selection conditions are made during the experiment such as changing or adding additional drugs in the Kishony experiment or adding additional selection pressures such as thermal stress in the Lenski experiment will slow the evolutionary process or possibly drive the populations to extinction.
Kleinman writes:
If they were to change their environments at any time during the performance of the experiment, that would mean they would be changing the evolutionary trajectories for increasing fitness making those evolutionary trajectories more complex.
Taq writes:
Why would it be more complex? If you changed temperature or salt concentrations at the same time I don't see how it would be more complex, as long as those conditions weren't immediately lethal.
I'll explain why the evolutionary trajectory is more complex when evolution must occur to simultaneous selection pressures using the Kishony experiment because it is easier to visualize.
Let's say that Kishony uses two drugs instead of one in his experiment. Let's say one drug requires mutations A1, A2, A3 for adaptation, and the other drug requires mutations B1, B2, B3. The 1,2 and 3 indicate drug regions 1,2 and 3 respectively. For a member of the wild-type to grow in the first region, that member needs mutations A1 and B1 (note that in the single drug experiment, A1 or B1 alone would allow growth in the first region depending on the drug).
The wild-type now starts to replicate in the drug-free region and when the colony size reaches about a billion, there is a reasonable probability of an A1 variant occurring and a B1 variant occurring. As the colony is growing the subset with the A1 and B1 mutations are growing as well but the vast majority of the population will still be wild-type (without either A1 or B1). In about 30 more doublings of the colony, the A1 and B1 subsets will achieve a billion replications and the A1 variant will have a reasonable probability of getting a B1 mutation and the B1 variant will have a reasonable probability of getting an A1 mutation. But, at the same time, the wild-type subset which had a population size of about a billion when the A1 and B1 variants appeared will also have done 30 more doublings (if the petri dish is large enough to allow this). Each different selection condition has its own particular set of mutations which might give improved fitness. Why would you think that adaptation to salt concentrations would give the require the same mutations as thermal stress? They don't. But if you want a population that adapts to both salt concentrations and thermal stress, a lineage needs to accumulate the mutations necessary for improved fitness and adaptation to both selection pressures. And that's going to take exponentially more replications than the single selection pressure environments.
If you want to see how to do the mathematics for adaptation to multiple simultaneous selection pressures, read this paper:
Just a moment...

This message is a reply to:
 Message 34 by Taq, posted 05-26-2020 5:35 PM Taq has replied

Replies to this message:
 Message 40 by Taq, posted 05-27-2020 11:59 AM Kleinman has replied

  
Kleinman
Member (Idle past 325 days)
Posts: 2142
From: United States
Joined: 10-06-2016


Message 38 of 239 (876744)
05-27-2020 9:28 AM


Responsed to nwr and jar
Kleinman writes:
The random trial is a replication and the two possible outcomes are, does a beneficial mutation occur or does a beneficial mutation does not occur.
nwr writes:
That's too simplistic.
A mutation might allow an organism to move to a slightly different environmental niche. The mutation could be detrimental in the current niche, but advantageous in the new niche.
Sure, that's just a different variant on a different evolutionary trajectory. That's why when Kishony runs his experiment with ciprofloxacin, any variant with a mutation that might be beneficial for trimethoprim (or any other class of antibiotics for that matter) cannot successfully evolve on his plate. There's not enough carrying capacity. But the mathematics for DNA evolution for each of these evolutionary trajectories is the same. It is all a matter of being able to replicate sufficiently to accumulate the mutations necessary to adapt to the selection pressure(s) of a given environment.
Kleinman writes:
Withholding antibiotics from someone who needs them not only has health consequences but economic consequences as well.
nwr writes:
I'm not aware of any serious suggestions to do that. More commonly the concern is with indiscriminate use of antibiotics, or the use of antibiotics in farm animal feed because it helps to fatten the animals.
You should be aware of this problem. The way it happens is a patient presents to the medical provider with relatively mild symptoms, eg, clear nasal discharge, and an irritated throat. The medical provider makes a clinical judgment and tells the patient it's probably allergies or a virus and to go home and take an allergy pill or some chicken soup. The medical provider doesn't do a close follow-up or instruct the patient correctly what to do if worse and the patient ends up with a more severe bacterial infection. This is actually a common problem as indicated by the hospital admission data (septicemia and pneumonia numbers 1 and 4 respectively).
With respect to the indiscriminate use of antibiotics, the cliche, "you can't treat viral infections with antibacterial drugs", is not quite accurate. The patients that got influenza in the H1N1 epidemic about 10 years ago had about a 60% co-infection rate with bacteria and it won't surprise me if when the data is analyzed from the covid epidemic that many of the deaths are associated with bacterial co-infections. That's the reason I think they were seeing some improvement when using azithromycin for treatment. It's also possible that azithromycin has some antiviral activity.
And the use of antibiotics may have been done in food-lot raised animals that are fed starch because they don't need the cellulose digesting bacteria in their rumen but giving antibiotics to range-fed animals will prevent them from digesting grass. But, let's say that antibiotics are given to corn or grain-fed feed-lot animals, what difference does that make to people getting drug-resistant infections? We've already shown that you don't need antibiotics for the drug-resistant variants to appear. We already have these drug-resistant variants in our gut, on our skin, and on the lining of our respiratory tract. It's the misunderstanding of the mathematics of evolution which leads to incorrect usage of antibiotics and the selection of these drug-resistant variants causing these multi-drug resistant infections.
jar writes:
Or in patients that start to feel better and so stop taking the full prescribed regimen.
That's correct. The patient may still have mild residual infection and not comprehend the signs. Many physicians tell their patients, "just take all 10 days of your antibiotics" is an inadequate way of instructing patients on the usage of antibiotic. Patients need to understand what they should expect from correct antimicrobial treatment since they are the ones doing the observation of their response to treatment. I give my patients 3 rules to follow when using antibiotics. First rule, if you don't see improvement by day 2 (reduced fever, reduced pain, a sense of better well being, etc.), call me. Second rule, if antibiotic is working, stay on treatment until completely better plus 3 days (to account for assymptomatic residual infection). And third, when stopping the antibiotic, watch for relapse, if that occurs, restart antibiotics and notify me.

Replies to this message:
 Message 39 by Taq, posted 05-27-2020 11:47 AM Kleinman has not replied

  
Kleinman
Member (Idle past 325 days)
Posts: 2142
From: United States
Joined: 10-06-2016


Message 41 of 239 (876752)
05-27-2020 1:15 PM
Reply to: Message 40 by Taq
05-27-2020 11:59 AM


Re: Trying to give nwr more than a vague understanding of DNA evolution
Kleinman writes:
That's correct that a mutation can be neutral or detrimental in one environment (what you call a genetic background) and beneficial in another.
Taq writes:
Genetic background and environment are two different things. What I am talking about is epistasis:
"Epistasis is a phenomenon in genetics in which the effect of a gene mutation is dependent on the presence or absence of mutations in one or more other genes, respectively termed modifier genes."
Epistasis - Wikipedia
From your wikipedia link writes:
Understanding of epistasis has changed considerably through the history of genetics and so too has the use of the term. In early models of natural selection devised in the early 20th century, each gene was considered to make its own characteristic contribution to fitness, against an average background of other genes. Some introductory courses still teach population genetics this way. Because of the way that the science of population genetics was developed, evolutionary geneticists have tended to think of epistasis as the exception. However, in general, the expression of any one allele depends in a complicated way on many other alleles.
You are conflating two concepts, the creation of new alleles (DNA evolution) and the expression of any one allele (epistasis).
Kleinman writes:
What changes could you make to the Kishony experiment to make his populations evolve more quickly?
Taq writes:
You could change growth temperature, salt concentration, protein sources, carbon sources, and so on. Growing bacteria in the same environment for long periods of time moves them towards a peak in the fitness landscape where very few if any mutations improving fitness.
Fitness landscape - Wikipedia
There is no indication that Kishony's team is running their experiment at non-optimal temperatures, non-optimal salt concentrations, non-optimal protein and carbon sources or any other selection pressure.
Spatiotemporal microbial evolution on antibiotic landscapes - PMC
But even if they did, why would you think that it would take fewer than a billion replications for each evolutionary step to evolve and adapt to the antibiotic selection pressure?
Kleinman writes:
Let's say that Kishony uses two drugs instead of one in his experiment. Let's say one drug requires mutations A1, A2, A3 for adaptation, and the other drug requires mutations B1, B2, B3.
For a single drug it would be A1, A2, and A3. It would be the same pathway, and it wouldn't be more complex. The pathways for resistance to each drug are the same no matter how many drugs are present.
Why would you think that adaptation to salt concentrations would give the require the same mutations as thermal stress?
Taq writes:
I never said they would. If the bacteria are adapting to multiple new challenges at the same time then this will select for many new beneficial mutations in many genes. If we are measuring evolution as the fixation of beneficial mutations, then a big change in environment will result in a higher rate of evolution.
The key point that you are missing here is that all these beneficial mutation somehow have to end up in some common lineage for this lineage to be adapted to all these selection pressures. It doesn't help one lineage which has a beneficial mutation for a given selection pressure if a different lineage gets a beneficial mutation for a different selection pressure. Putting it into the context of the Kishony experiment, a beneficial mutation for trimethoprim doesn't help that variant when the selection pressure is ciprofloxacin. But when Kishony runs his experiment with two drugs, only when some member of the population has a beneficial mutation for both ciprofloxacin and for trimethoprim will that member be able to grow in the next higher drug concentration region.
And fixation is not required for DNA evolution to occur. If you think it does, why are the drug sensative variants still growing when the drug resistant variants have evolved? Competition and fixation only slow the DNA evolution process.
Taq writes:
Also, there could be overlap in salt and temperature adaptations, such as chaperone proteins that could stabilize proteins in both high temps and different salt concentrations (i.e. heat shock proteins).
You are speculating. Try understanding these simple evolutionary experiments. The math is not that hard. Even your slot-machine analogy is not that bad. The math really isn't much different than the binomial probability problem. When you pull on the arm of the one-armed bandit, you may have tiny chance of winning the big jackpot and you might have a better chance of winning a smaller jackpot but the joint probability of winning both has to be computed using the multiplication rule. (Of course, you won't win both on a single pull, the events are mutually exclusive).

This message is a reply to:
 Message 40 by Taq, posted 05-27-2020 11:59 AM Taq has replied

Replies to this message:
 Message 42 by Taq, posted 05-27-2020 1:34 PM Kleinman has replied

  
Newer Topic | Older Topic
Jump to:


Copyright 2001-2023 by EvC Forum, All Rights Reserved

™ Version 4.2
Innovative software from Qwixotic © 2024