Register | Sign In


Understanding through Discussion


EvC Forum active members: 65 (9162 total)
2 online now:
Newest Member: popoi
Post Volume: Total: 915,819 Year: 3,076/9,624 Month: 921/1,588 Week: 104/223 Day: 2/13 Hour: 1/0


Thread  Details

Email This Thread
Newer Topic | Older Topic
  
Author Topic:   Do you really understand the mathematics of evolution?
PaulK
Member
Posts: 17822
Joined: 01-10-2003
Member Rating: 2.2


Message 215 of 239 (878566)
07-01-2020 4:05 PM
Reply to: Message 214 by Kleinman
07-01-2020 2:01 PM


Re: An important note regarding models of evolution
quote:
I'm not the one who thinks you can do phylogenetic analysis with this model, especially when you think you tell how many generations separate two genetic sequences.
Since your preferred model would be worse for that, then your opinion doesn’t count for much. You can’t get a decent model unless you understand what you are modelling.
quote:
Sure, I'm confused about your explanation, first, you say this model is about neutral evolution, but then you say it doesn't compute fixation, it's about the frequencies of bases being equally distributed and it happen long ago. What kind of information can you get out of this model? Can you put in a portion of your genome and a portion of the genome from a banana and tell how many generations back your most recent common ancestor is?
Yes, you are indeed confused due to your failure to understand what you are talking about. With a sufficiently large portion of the genome - and with decent measures of the rate of change for those portions it should certainly be possible to come up with a rough estimate.

This message is a reply to:
 Message 214 by Kleinman, posted 07-01-2020 2:01 PM Kleinman has replied

Replies to this message:
 Message 216 by Kleinman, posted 07-01-2020 4:58 PM PaulK has replied

  
PaulK
Member
Posts: 17822
Joined: 01-10-2003
Member Rating: 2.2


Message 217 of 239 (878587)
07-02-2020 1:52 AM
Reply to: Message 216 by Kleinman
07-01-2020 4:58 PM


Re: An important note regarding models of evolution
quote:
My model explains how DNA evolution works and you could do the same with a Markov model if you understood how to write the correct transition matrix.
You say that it does, but the basis of the claim is that it handles an extremely atypical case. You claim to have been an engineer so you must understand how wrong that is.
quote:
Sure, but you aren't going to do it by cherrypicking the genome for tiny homologous portions.
Sure, the strawman you invented isn’t going to work.
quote:
In fact, you have to use the entire genome and account for all the differences
Statisticians would disagree. Proper sampling should be quite adequate.
quote:
You can't even use the entire coding portion of the genome because even if the coding genes are similar between two species, it's the non-coding regulatory portion of the genome (what the fish evolve into mammals clique like to call junk DNA) is what makes creatures what they are and that's a much, much larger portion of the genome than the coding portion.
While there is more non-coding DNA than coding DNA the regulatory portion is only a small part of that. And I don’t like to call regulatory DNA junk. You do like making things up.
Genuine junk - which excludes regulatory regions - is worth looking at because it is not under any selective constraint,
quote:
It only takes a mere 50 million generations to reach equilibrium. Do you want to explain that to us, Professor Corey?
What’s to explain? 50,000,000 generations is a long time in most vertebrates. Since you have an interest in bird evolution the fact that most of the birds I am familiar with have only one generation a year would seem to be relevant

This message is a reply to:
 Message 216 by Kleinman, posted 07-01-2020 4:58 PM Kleinman has replied

Replies to this message:
 Message 218 by Kleinman, posted 07-02-2020 5:41 AM PaulK has replied

  
PaulK
Member
Posts: 17822
Joined: 01-10-2003
Member Rating: 2.2


Message 219 of 239 (878617)
07-02-2020 11:54 AM
Reply to: Message 218 by Kleinman
07-02-2020 5:41 AM


Re: An important note regarding models of evolution
I think that replying to this part is all that needs to be said:
quote:
Kleinman writes:
Sure, but you aren't going to do it by cherrypicking the genome for tiny homologous portions.
PaulK writes:
Sure, the strawman you invented isn’t going to work.
You are a liar PaulK, this is exactly how the fish evolves into mammals clique uses these Markov models. That's why you won't present any real examples of how these models are used.
You call me a liar but you don’t say anything that backs up your accusation. Show us evidence of actual cherry picking or admit that you haven’t got any.

This message is a reply to:
 Message 218 by Kleinman, posted 07-02-2020 5:41 AM Kleinman has replied

Replies to this message:
 Message 220 by Kleinman, posted 07-02-2020 12:47 PM PaulK has replied

  
PaulK
Member
Posts: 17822
Joined: 01-10-2003
Member Rating: 2.2


Message 221 of 239 (878634)
07-02-2020 1:22 PM
Reply to: Message 220 by Kleinman
07-02-2020 12:47 PM


Re: An important note regarding models of evolution
quote:
If there was a league of lazy dumb asses, you would be on the all-star team
I’d refuse to play in any team that had you as the captain.
quote:
You definitely don't want to pick any sour cherries. So, son of banana parents, that's how the cherry-picking should be done if you want to show that your parents are bananas.
The obvious problem with the quotes is that the sequences are picked before the alignment. So, the choice of sequences isn’t cherry picked. The alignment is constrained by that choice, and I doubt you’d get a good match by pure chance. Even with only two sequences.
It seems to me that alignment is needed to handle insertions and deletions. If you just try matching bases to bases without taking that into account you’ll hit problems.

This message is a reply to:
 Message 220 by Kleinman, posted 07-02-2020 12:47 PM Kleinman has replied

Replies to this message:
 Message 222 by Kleinman, posted 07-02-2020 1:39 PM PaulK has replied

  
PaulK
Member
Posts: 17822
Joined: 01-10-2003
Member Rating: 2.2


Message 223 of 239 (878636)
07-02-2020 2:05 PM
Reply to: Message 222 by Kleinman
07-02-2020 1:39 PM


Re: An important note regarding models of evolution
quote:
Here's the contact information for the authors of the page, take it up with them:
The Tech Interactive
201 S. Market St.
San Jose, CA 95113
1-408-294-8324
That’s the contact information for the publishers, not the author. Unlike you, I’m not too lazy to find out that much.
quote:
I'm sure they are sitting by the phone right now waiting for a call from the king explaining how frameshift mutations have to be taken into account in their phylogenic analysis
Since the author insists on alignment which does take account of frame shifting, they hardly need to be told. It’s only the guy who think that aligning the sequences that have already been chosen is cherry picking the sequence that needs telling. And I already told him.
So chalk up yet another Kleinman blunder.

This message is a reply to:
 Message 222 by Kleinman, posted 07-02-2020 1:39 PM Kleinman has replied

Replies to this message:
 Message 224 by Kleinman, posted 07-02-2020 2:24 PM PaulK has replied

  
PaulK
Member
Posts: 17822
Joined: 01-10-2003
Member Rating: 2.2


Message 225 of 239 (878652)
07-02-2020 4:39 PM
Reply to: Message 224 by Kleinman
07-02-2020 2:24 PM


Re: An important note regarding models of evolution
quote:
Maybe the king can find the contact information by reading some fossil tea-leaves.
I could, but I don’t need it. She didn’t support your claim of cherry-picking and that’s all that needs to be said.
quote:
The king will now explain how fish evolve into mammals with frameshift mutations.
Poor Kleinman. He thinks that if we acknowledge that frameshift mutations - and indels that don’t shift the frame - we must assume that all evolutionary change is explained by frameshift mutations.
quote:
What really scares me is that I think PaulK actually was a biology teacher.
Well there’s your imagination running away with you again. I’ve never claimed any qualifications in biology and I’ve never worked as a teacher.

This message is a reply to:
 Message 224 by Kleinman, posted 07-02-2020 2:24 PM Kleinman has replied

Replies to this message:
 Message 226 by Kleinman, posted 07-02-2020 5:06 PM PaulK has replied

  
PaulK
Member
Posts: 17822
Joined: 01-10-2003
Member Rating: 2.2


Message 227 of 239 (878669)
07-03-2020 12:50 AM
Reply to: Message 226 by Kleinman
07-02-2020 5:06 PM


Re: An important note regarding models of evolution
quote:
And you in the fish evolve into mammals clique think I have trouble with the english language.
Oh, you have plenty of problems, but that’s not one on my list.
quote:
What do you think that "Make sure you’re comparing the same gene!" means ding-dong? I really hope you are only pretending to be a dummy.
Obviously you have to compare the same gene. Do you really think you can compare a gene taken from one species with a random chunk of DNA taken from another and get a decent phylogeny? You must be hoping that I am stupid enough to fall for these silly deceptions.

This message is a reply to:
 Message 226 by Kleinman, posted 07-02-2020 5:06 PM Kleinman has replied

Replies to this message:
 Message 228 by Kleinman, posted 07-03-2020 5:48 AM PaulK has replied

  
PaulK
Member
Posts: 17822
Joined: 01-10-2003
Member Rating: 2.2


Message 229 of 239 (878685)
07-03-2020 9:01 AM
Reply to: Message 228 by Kleinman
07-03-2020 5:48 AM


Re: An important note regarding models of evolution
quote:
So, according to Jukes-Cantor calculation, if you compare a single gene from two different species with a single base difference between them and a mutation rate of e-8, you get 50,000,000 generations separating the two species.
No. According to the Jukes-Cantor calculation 3/4 of the bases should have changed in 50,000,000 generations. That’s going to be more than one. A lot more.
quote:
So, ding-dong, what if you compare a collection of 10 equivalent genes at the same time, each with only a single base difference, that means 500,000,000 generations separating the two species
Assuming that the genes were of roughly equal length it would mean that they agreed on the divergence time. Which would be well under. 50,000,000 generations.
If you can handle the math, or even read the pages you cite you should know that much.

This message is a reply to:
 Message 228 by Kleinman, posted 07-03-2020 5:48 AM Kleinman has replied

Replies to this message:
 Message 230 by Kleinman, posted 07-03-2020 9:34 AM PaulK has replied

  
PaulK
Member
Posts: 17822
Joined: 01-10-2003
Member Rating: 2.2


Message 231 of 239 (878687)
07-03-2020 9:47 AM
Reply to: Message 230 by Kleinman
07-03-2020 9:34 AM


Re: An important note regarding models of evolution
quote:
So, when you said "I’ve never claimed any qualifications in biology", you are now claiming some qualifications? So, which 3/4ths of bases have changed and show us how you use that information for constructing a DNA phylogenetic tree.
I don’t claim any biology qualifications. But this is mathematics. And no, Jukes-Cantor won’t tell you which bases change as you ought to know.
quote:
So, as the number of genes being analyzed increases and the number of genetic differences increases, the divergence time goes down.
Wrong again. The divergence time is lower than your estimate because your estimate is hopelessly wrong. And since the divergence is based on the proportion of changed bases, it won’t change unless that proportion is changed.

This message is a reply to:
 Message 230 by Kleinman, posted 07-03-2020 9:34 AM Kleinman has replied

Replies to this message:
 Message 232 by Kleinman, posted 07-03-2020 10:03 AM PaulK has replied

  
PaulK
Member
Posts: 17822
Joined: 01-10-2003
Member Rating: 2.2


Message 233 of 239 (878690)
07-03-2020 10:18 AM
Reply to: Message 232 by Kleinman
07-03-2020 10:03 AM


Re: An important note regarding models of evolution
quote:
Are you now claiming some qualifications in mathematics? If so, show us the mathematics used to compute the divergence time if two sites are considered.
No sensible person would use only two sites. You couldn’t get a usable estimate that way.
quote:
You just said, "Assuming that the genes were of roughly equal length it would mean that they agreed on the divergence time. Which would be well under. 50,000,000 generations." and now you are claiming qualifications in mathematics, show us the mathematics that it would be "well under. 50,000,000 generations".
I said that they would agree on the divergence time. I.e. each considered individually would give about the same divergence time.
I also said that the divergence time would be well under 50,000,000 generations because only a small proportion of the bases has changed. In 50,000,000 generations 0.75 of the bases would be changed according to Jukes-Cantor with the given parameters.

This message is a reply to:
 Message 232 by Kleinman, posted 07-03-2020 10:03 AM Kleinman has replied

Replies to this message:
 Message 234 by Kleinman, posted 07-03-2020 10:38 AM PaulK has replied

  
PaulK
Member
Posts: 17822
Joined: 01-10-2003
Member Rating: 2.2


Message 235 of 239 (878703)
07-03-2020 10:54 AM
Reply to: Message 234 by Kleinman
07-03-2020 10:38 AM


Re: An important note regarding models of evolution
quote:
Do the math for any number of sites you think is sensible. And please show your math.
Let’s start with some simpler math. When the Jukes-Cantor model reaches equilibrium, 3 quarters of the bases should have changed.
If at that point only one base in the gene has changed, how many bases would you expect to be in the gene?
And for a bonus biology question: could a gene be that length?
quote:
So the divergence of two sites would give "about the same divergence time"?
Well let’s try some more simple mathematics. If one base in a gene 10,000 bases long has changed what proportion of bases has changed ?
If you have two genes, each 10,000 bases long and one base in each has changed, what proportion of bases has changed?
Are the two proportions different ?
quote:
You just said, "No sensible person would use only two sites. You couldn’t get a usable estimate that way." Why can you get divergence times for one site or many sites but two sites is not sensible?
Because you aren’t just counting the sites that have changed, you’re also counting all the sites that haven’t. Which comes to a lot more than just two.

This message is a reply to:
 Message 234 by Kleinman, posted 07-03-2020 10:38 AM Kleinman has replied

Replies to this message:
 Message 236 by Kleinman, posted 07-03-2020 11:59 AM PaulK has replied

  
PaulK
Member
Posts: 17822
Joined: 01-10-2003
Member Rating: 2.2


Message 237 of 239 (878706)
07-03-2020 12:22 PM
Reply to: Message 236 by Kleinman
07-03-2020 11:59 AM


Re: An important note regarding models of evolution
quote:
Why couldn't all the bases have changed? After all, it is 50,000,000 generations. Can't some member of that population over 50,000,000 generations get a mutation at any site in the genome?
Feel free to calculate the probability of that happening. 0.75^n where n is the number of bases.
quote:
Why couldn't all the bases have changed? After all, it is 50,000,000 generations. Can't some member of that population over 50,000,000 generations get a mutation at any site in the genome?
I’m sure you know that you claimed that only one base would change in a gene in 50,000,000 generations. So the expected length of the gene would be 1/0.75 bases. Let’s generously round up and say two. Not exactly plausible is it?
And of course, if you really understood the mathematics you would know that already.
So obviously you are either a fraud or a troll. And either way you need help.

This message is a reply to:
 Message 236 by Kleinman, posted 07-03-2020 11:59 AM Kleinman has replied

Replies to this message:
 Message 238 by Kleinman, posted 07-03-2020 12:45 PM PaulK has not replied

  
Newer Topic | Older Topic
Jump to:


Copyright 2001-2023 by EvC Forum, All Rights Reserved

™ Version 4.2
Innovative software from Qwixotic © 2024