Gene frequencies and evolution

I belive genes cooperate with other genes in groups for large evolutionary changes.

Mats Liljedahl

phillip smith said:

Evolution has been described as changes in gene frequencies. I have myself espoused this point of
view. I am beginning to have some doubts. The gene centered view of selection with coefficients for
each allele where the differences of genomic back ground are supposed to average out and the
frequency of the gene changes because of the mean selective advantage it confers to individuals
bearing that allele. Please ignore genetic drift for the purpose of this argument. Does any one
know of any theoretical discussion of this assertion. I can think of three confounding factors for
such as argument.
1) Epistatic interactions between the rest of the genome on the coefficent of selection are unknown
and perhaps unknowable
2) selection events are insufficiently uniform. I.e every individuals death is a unique event
3) the combination of the above prevents the normalisation of the data to quantify the effect of a
particular allele at a particular locus.

I would have though fisher and others may have dealt with these issues any one know of any
investigations in to this area

Thanks n advance

phillip smith said:

Evolution has been described as changes in gene frequencies. I have myself espoused this point of
view. I am beginning to have some doubts. The gene centered view of selection with coefficients
for each allele where the differences of genomic back ground are supposed to average out and the
frequency of the gene changes because of the mean selective advantage it confers to individuals
bearing that allele. Please ignore genetic drift for the purpose of this argument. Does any one
know of any theoretical discussion of this assertion.

Fisher 1930? Or do you want something readable?

[moderator's note: Hey! I think Fisher 1930 is perfectly readable, and since it's available in a
Dover paperback edition, it's really cheap, too -- certainly worth tackling. Besides, if _I_ had to
wade through it, YOU should too. - JAH]

You should try a good textbook on quantitative genetics. Either Falconer & McKay or Lynch & Walsh
would be a good start (L&W is big, though)

Quoted message said:

I can think of three confounding factors for such as argument.
1) Epistatic interactions between the rest of the genome on the coefficent of selection are
unknown and perhaps unknowable

This is a only problem if there is also linkage disequilibrium.

Also note that for a lot of the genome, the genes are not polymorphic (=no variation), and most of
the rest of the genome probably have negligible epistatic effects.

So, at least the problem is smaller. :-)

Quoted message said:

2) selection events are insufficiently uniform. I.e every individuals death is a unique event

But that brings you back to drift. With an infinite number of individual deaths, the other effects
average out. With a finite number, the difference is drift.

Quoted message said:

3) the combination of the above prevents the normalisation of the data to quantify the effect of a
particular allele at a particular locus.

No, the first problem was solved by Fisher (but the modern formulation is easier to follow),
basically you use the additive effect of an allele. See a textbook for the details of what
additive means.

Quoted message said:

I would have though fisher and others may have dealt with these issues any one know of any
investigations in to this area

Yep. Fisher looked at both, in large populations.

Bob

--
Bob O'Hara

Rolf Nevanlinna Institute
P.O. Box 4 (Yliopistonkatu 5) FIN-00014 University of Helsinki Finland Telephone: +358-9-191 23743
Mobile: +358 50 599 0540 Fax: +358-9-191 22 779 WWW: rni.helsinki.fi~bohOpen ↗

Anon. said:

phillip smith said:

Evolution has been described as changes in gene frequencies. I have myself espoused this point
of view. I am beginning to have some doubts. The gene centered view of selection with
coefficients for each allele where the differences of genomic back ground are supposed to
average out and the frequency of the gene changes because of the mean selective advantage it
confers to individuals bearing that allele. Please ignore genetic drift for the purpose of this
argument. Does any one know of any theoretical discussion of this assertion.

Fisher 1930? Or do you want something readable?

[moderator's note: Hey! I think Fisher 1930 is perfectly readable, and since it's available in a
Dover paperback edition, it's really cheap, too -- certainly worth tackling. Besides, if _I_ had
to wade through it, YOU should too. - JAH]

You should try a good textbook on quantitative genetics. Either Falconer & McKay or Lynch & Walsh
would be a good start (L&W is big, though)

Quantitative genetics make a lot of simplifications in order to be able to formulate hypotheses that
they can manage to test. It shouldn't surprise anyone that the assumptions are not what is observed
in nature. All you can do is ask if it is a close enough approximation so that it can tell you
something about the system.

Quoted message said:

Quoted message said:

I can think of three confounding factors for such as argument.
1) Epistatic interactions between the rest of the genome on the coefficent of selection are
unknown and perhaps unknowable

This is a only problem if there is also linkage disequilibrium.

Also note that for a lot of the genome, the genes are not polymorphic (=no variation), and most of
the rest of the genome probably have negligible epistatic effects.

So, at least the problem is smaller. :-)

Epistasis in quantitative genetics is just any kind of gene interaction. This just alters the
additive effect of any given allele that is involved in the interactions. If the additive effect
goes to zero you can't select for that allele and it would be neutral. I doubt that this happens
very often.

Quoted message said:

Quoted message said:

2) selection events are insufficiently uniform. I.e every individuals death is a unique event

But that brings you back to drift. With an infinite number of individual deaths, the other effects
average out. With a finite number, the difference is drift.

Quoted message said:

3) the combination of the above prevents the normalisation of the data to quantify the effect of
a particular allele at a particular locus.

No, the first problem was solved by Fisher (but the modern formulation is easier to follow),
basically you use the additive effect of an allele. See a textbook for the details of what
additive means.

Quoted message said:

I would have though fisher and others may have dealt with these issues any one know of any
investigations in to this area

Yep. Fisher looked at both, in large populations.

Bob

The reservations of the first poster seems to be about the accuracy of determination of allele
effects. Epistasis will often reduce the rate of change of the allele frequency due to
selection, and random events can dramatically change allele frequencies in small populations.
Couple them together and they will likely increase the frequency where the less advantagous
allele is fixed by "accident" in a population, but it would still have to occur by accident and
against selection pressure.

Ron Okimoto

in article [email hidden], Anon. at
[email hidden] wrote on 20/1/04 6:01 AM:

Quoted message said:

Fisher 1930? Or do you want something readable?

[moderator's note: Hey! I think Fisher 1930 is perfectly readable, and since it's available in a
Dover paperback edition, it's really cheap, too -- certainly worth tackling. Besides, if _I_ had
to wade through it, YOU should too. - JAH]

You should try a good textbook on quantitative genetics. Either Falconer & McKay or Lynch & Walsh
would be a good start (L&W is big, though)

Thanks for those pointers. I have on many occasion tried to wade through Crow and Kimura and CC Li

Quoted message said:

Quoted message said:

I can think of three confounding factors for such as argument.
1) Epistatic interactions between the rest of the genome on the coefficent of selection are
unknown and perhaps unknowable

This is a only problem if there is also linkage disequilibrium.

Also note that for a lot of the genome, the genes are not polymorphic (=no variation), and most of
the rest of the genome probably have negligible epistatic effects.

Jus checking for data on the web I found some data on heterozygosity in natural populations Medfly
about 8% with estimates of heterozygosity fluctuating substantially around that in a few studies I
could find. WE probably wont have decent figure till population genomics becomes technically
possible. But certainly figures above 1 % would not be supprising. I think the human genome is
supposed to be around 20,000 genes so around so we might say you have to ignore epistatic
interations of 200 genes on the locus you are investigating. That sounds like a lot tome. Especially
since you are already ignoring all the non coding DNA ad any epignetic effects

Quoted message said:

So, at least the problem is smaller. :-)

Or perhaps just underestimated

Quoted message said:

Quoted message said:

2) selection events are insufficiently uniform. I.e every individuals death is a unique event

But that brings you back to drift. With an infinite number of individual deaths, the other effects
average out. With a finite number, the difference is drift.

Fair enough. What if no two deaths are the same? I.e same stage of development, same cause,
same genome.

Quoted message said:

Quoted message said:

3) the combination of the above prevents the normalisation of the data to quantify the effect of
a particular allele at a particular locus.

No, the first problem was solved by Fisher (but the modern formulation is easier to follow),
basically you use the additive effect of an allele. See a textbook for the details of what
additive means.

LOL We won't go there.

Quoted message said:

Quoted message said:

I would have though fisher and others may have dealt with these issues any one know of any
investigations in to this area

Yep. Fisher looked at both, in large populations.

I think I tried to wade through it a few years. I from memory he used a term h for epistasis but
may be it was some one else

--

Phillip Smith phills@(buggger).co.nz replace bugger with ihug applied-evolution.co.nzapplied-evolution.co.nzOpen ↗

"he who is smeared with blubber has the kindest heart" -- a Greenland Eskimo adage

phillip smith said:

in article [email hidden], Anon. at [email hidden]
wrote on 20/1/04 6:01 AM:

Quoted message said:

Fisher 1930? Or do you want something readable?

[moderator's note: Hey! I think Fisher 1930 is perfectly readable, and since it's available in a
Dover paperback edition, it's really cheap, too -- certainly worth tackling. Besides, if _I_ had
to wade through it, YOU should too. - JAH]

You should try a good textbook on quantitative genetics. Either Falconer & McKay or Lynch & Walsh
would be a good start (L&W is big, though)

Thanks for those pointers. I have on many occasion tried to wade through Crow and Kimura and CC Li

Quoted message said:

Quoted message said:

I can think of three confounding factors for such as argument.
1) Epistatic interactions between the rest of the genome on the coefficent of selection are
unknown and perhaps unknowable

This is a only problem if there is also linkage disequilibrium.

Also note that for a lot of the genome, the genes are not polymorphic (=no variation), and most of
the rest of the genome probably have negligible epistatic effects.

Jus checking for data on the web I found some data on heterozygosity in natural populations Medfly
about 8% with estimates of heterozygosity fluctuating substantially around that in a few studies I
could find. WE probably wont have decent figure till population genomics becomes technically
possible. But certainly figures above 1 % would not be supprising. I think the human genome is
supposed to be around 20,000 genes so around so we might say you have to ignore epistatic
interations of 200 genes on the locus you are investigating. That sounds like a lot tome.
Especially since you are already ignoring all the non coding DNA ad any epignetic effects

Quoted message said:

So, at least the problem is smaller. :-)

Or perhaps just underestimated

Quoted message said:

Quoted message said:

2) selection events are insufficiently uniform. I.e every individuals death is a unique event

But that brings you back to drift. With an infinite number of individual deaths, the other effects
average out. With a finite number, the difference is drift.

Fair enough. What if no two deaths are the same? I.e same stage of development, same cause,
same genome.

Well, that can only happen in a finite population. :-)

In one sense, no two deaths are the same, as every individual is unique. What is important is the
way that an organism's genes affect the probabilities of the different deaths (and othe events, of
course), so you can average over them (or, for calculating fitness, take the expectation).

Bob

--
Bob O'Hara

Dept. of Mathematics and Statistics
P.O. Box 4 (Yliopistonkatu 5) FIN-00014 University of Helsinki Finland Telephone: +358-9-191 23743
Mobile: +358 50 599 0540 Fax: +358-9-191 22 779 WWW: rni.helsinki.fi~bohOpen ↗

AA:- Mats Liljedahl quote: I belive genes cooperate with other genes in groups for large
evolutionary changes.

PS:

Quoted message said:

Evolution has been described as changes in gene frequencies. I have myself espoused this point of
view. I am beginning to have some doubts. The gene centered view of selection with coefficients for
each allele where the differences of genomic back ground are supposed to average out and the
frequency of the gene changes because of the mean selective advantage it confers to individuals
bearing that allele. Please ignore genetic drift for the purpose of this argument. Does any one
know of any theoretical discussion of this assertion.

JE:- The above description accurately describes Neo Darwinistic gene centric (heavily
oversimplified), models of Darwinian organism fitness. These models can be very useful and have
helped our understanding of Darwin's evolutionary process. However, they have been terribly misused.
Nobody here will discuss even just a possibility of model misuse. Trying to raise this topic within
sbe, is harder than raising the dead. The only thing that sometimes works are arguments heavily
spiced with rhetoric. It appears the only way to wake the slumbering giant of Neo Darwinism is to
attack its ego.

All the problems are epistemological:
1) The difference between a model and a theory (if any).
2) The validity of models and theories.
3) The testing of models and theories.
4) How and why one model, or one theory, can be allowed to replace another producing the evolution
of ideas within the sciences.

Quoted message said:

AA:- I can think of three confounding factors for such as argument.
1) Epistatic interactions between the rest of the genome on the coefficent of selection are unknown
and perhaps unknowable

JE:- Genes are selected together and not separately. The organism always represents just ONE genomic
fitness, i.e. all the genes in the genome are selected at just the _one_, same level: the fertile
organism level of selection. This traditional level of selection represents the _first_ independent
level of fitness within nature, both logically and empirically. Independence is represented
mathematically by the term "addition". The fertile organism level is the first additive fitness
level than can be measured within nature. Selection is determined by what happens at this first
independent level of selection. Subsequent additive levels represent group selective logic. Group
selection fails because only the first additive fitness level is effective. Subsequent levels can
only go with and not against, selection at the first independent level of selection. Thus fitness
multilevels are all complimentary to the 1st additive fitness level. This simple logic is not
understood within Neo Darwinism.

Hamilton's supposed independent gene level of selection does not exist in nature because non
additive gene fitness epistasis prohibits it. Non additive gene fitnesses must be selected together.
This fact denies any selective independence for a genomic gene. Thus models of independent gene
selection are massive oversimplifications of the known facts of nature. Allowing such a model to
compete and win against the theory it was simplified from, is an absurdity. Neo Darwinism is full of
such absurdities. These errors heavily discount the enormous contribution Neo Darwinistic models
can/have made to evolutionary theory. Until model misuse is tackled head on, the negatives generated
by model misuse outweigh the positives of correct model usage.

Quoted message said:

AA:-
2) selection events are insufficiently uniform. I.e every individuals

death

Quoted message said:

is a unique event.

JE:- Death is of less importance than reproducing fertile forms.

Quoted message said:

AA:-
3) the combination of the above prevents the normalisation of the data to quantify the effect of a
particular allele at a particular locus. I would have though fisher and others may have dealt
with these issues any one know of any investigations in to this area.

JE:- Individual alleles will be selected via the phenotype that they help code. Biologists
traditionally make good guesses as to which phenotype is more important. All that can be done at the
moment is attempt to predict gene freq, mostly retrospectively, from just a guess at a phenotype's
fitness where we have _some_ understanding of the genes that code for that phenotype. Note that
Fisher deleted gene fitness epistasis as "inherited" but "non heritable" and thus, "non selectable",
i.e. removed almost everything of biological substance.

Regards,

John Edser Independent Researcher

PO Box 266 Church Pt NSW 2105 Australia

[email hidden]

in article [email hidden], Ron Okimoto at
[email hidden] wrote on 21/1/04 5:39 PM:

Quoted message said:

Anon. said:

phillip smith said:

Evolution has been described as changes in gene frequencies. I have myself espoused this point
of view. I am beginning to have some doubts. The gene centered view of selection with
coefficients for each allele where the differences of genomic back ground are supposed to
average out and the frequency of the gene changes because of the mean selective advantage it
confers to individuals bearing that allele. Please ignore genetic drift for the purpose of this
argument. Does any one know of any theoretical discussion of this assertion.

Fisher 1930? Or do you want something readable?

[moderator's note: Hey! I think Fisher 1930 is perfectly readable, and since it's available in a
Dover paperback edition, it's really cheap, too -- certainly worth tackling. Besides, if _I_ had
to wade through it, YOU should too. - JAH]

You should try a good textbook on quantitative genetics. Either Falconer & McKay or Lynch & Walsh
would be a good start (L&W is big, though)

Quantitative genetics make a lot of simplifications in order to be able to formulate hypotheses
that they can manage to test. It shouldn't surprise anyone that the assumptions are not what is
observed in nature. All you can do is ask if it is a close enough approximation so that it can
tell you something about the system.

That is what I am asking

Quoted message said:

Quoted message said:

Quoted message said:

I can think of three confounding factors for such as argument.
1) Epistatic interactions between the rest of the genome on the coefficent of selection are
unknown and perhaps unknowable

This is a only problem if there is also linkage disequilibrium.

Also note that for a lot of the genome, the genes are not polymorphic (=no variation), and most
of the rest of the genome probably have negligible epistatic effects.

So, at least the problem is smaller. :-)

Epistasis in quantitative genetics is just any kind of gene interaction. This just alters the
additive effect of any given allele that is involved in the interactions. If the additive effect
goes to zero you can't select for that allele and it would be neutral. I doubt that this happens
very often.

Quoted message said:

Quoted message said:

2) selection events are insufficiently uniform. I.e every individuals death is a unique event

But that brings you back to drift. With an infinite number of individual deaths, the other
effects average out. With a finite number, the difference is drift.

Quoted message said:

3) the combination of the above prevents the normalisation of the data to quantify the effect of
a particular allele at a particular locus.

No, the first problem was solved by Fisher (but the modern formulation is easier to follow),
basically you use the additive effect of an allele. See a textbook for the details of what
additive means.

Quoted message said:

I would have though fisher and others may have dealt with these issues any one know of any
investigations in to this area

Yep. Fisher looked at both, in large populations.

Bob

The reservations of the first poster seems to be about the accuracy of determination of allele
effects. Epistasis will often reduce the rate of change of the allele frequency due to selection,
and random events can dramatically change allele frequencies in small populations. Couple them
together and they will likely increase the frequency where the less advantagous allele is fixed by
"accident" in a population, but it would still have to occur by accident and against selection
pressure.

Ron Okimoto

The question is. Is the definition of evolution as changes in gene frequencies reasonable. I
have no question that the frequency of phenotypic traits can change under selection. This not
the same thing.

In small populations it seems that drift is a factor in gene frequencies so we can say frequencies
change but we don't know why In large populations we can say that drift can be discounted. Providing
epistasis is ignored. If you take an infinite population then it is impossible to discount epistasis
as we have an infinite number of alleles at each locus. So this is not terriblu help full. WE might
say that the probability of epistasis is a function of population size and genome size. This seems
to suggest that changes in allele in frequencies under selection may be a rare event

--

Phillip Smith phills@(buggger).co.nz replace bugger with ihug applied-evolution.co.nzapplied-evolution.co.nzOpen ↗

"he who is smeared with blubber has the kindest heart" -- a Greenland Eskimo adage

phillip smith <[email hidden]> wrote in
:"]news:[email hidden]:

Quoted message said:

Evolution has been described as changes in gene frequencies. I have myself espoused this point of
view. I am beginning to have some doubts. The gene centered view of selection with coefficients
for each allele where the differences of genomic back ground are supposed to average out and the
frequency of the gene changes because of the mean selective advantage it confers to individuals
bearing that allele. Please ignore genetic drift for the purpose of this argument. Does any one
know of any theoretical discussion of this assertion. I can think of three confounding factors for
such as argument.
1) Epistatic interactions between the rest of the genome on the coefficent of selection are
unknown and perhaps unknowable

Exactly what difference does this make? When a tiger is chasing Joe and Fred, and Fred is slower,
the tiger is not going to go after Joe instead because the gene that makes Joe faster has epistatic
interactions.

Quoted message said:

2) selection events are insufficiently uniform. I.e every individuals death is a unique event

The question in general is not death, it is life. Many are born, few survive. This is one of the key
principles of evolution.

Quoted message said:

3) the combination of the above prevents the normalisation of the data to quantify the effect of a
particular allele at a particular locus.

Quoted message said:

I would have though fisher and others may have dealt with these issues any one know of any
investigations in to this area

Others have responded to this point.

But perhaps evolution is better thought of as changes in phenotype. We typically say that horseshoe
crabs are "unchanged" after millions of years, even though their gene frequencies today are almost
certainly significantly different from their gene frequencies millions of years ago.

Yours,

Bill Morse

phillip smith said:

in article [email hidden], Ron Okimoto at [email hidden] wrote on 21/1/04
5:39 PM:

<snip>

Quoted message said:

Quoted message said:

The reservations of the first poster seems to be about the accuracy of determination of allele
effects. Epistasis will often reduce the rate of change of the allele frequency due to selection,
and random events can dramatically change allele frequencies in small populations. Couple them
together and they will likely increase the frequency where the less advantagous allele is fixed by
"accident" in a population, but it would still have to occur by accident and against selection
pressure.

Ron Okimoto

The question is. Is the definition of evolution as changes in gene frequencies reasonable. I
have no question that the frequency of phenotypic traits can change under selection. This not
the same thing.

In small populations it seems that drift is a factor in gene frequencies so we can say frequencies
change but we don't know why In large populations we can say that drift can be discounted.
Providing epistasis is ignored. If you take an infinite population then it is impossible to
discount epistasis as we have an infinite number of alleles at each locus. So this is not terriblu
help full. WE might say that the probability of epistasis is a function of population size and
genome size. This seems to suggest that changes in allele in frequencies under selection may be a
rare event

But you need linkage disequilibrium (LD) as well as epistasis to change the effects of a single
gene. LD is reduced by recombination, so you'd need the interacting genes to be close to each other,
or for the epistasis to be large, for the affect to be appreciable.

Even if you do have evolution due to epistasis, this will probably entail changes in gene
frequencies anyway. It's just that you have to consider more than one gene to understand the
dynamics correctly.

Bob

--
Bob O'Hara

Dept. of Mathematics and Statistics
P.O. Box 4 (Yliopistonkatu 5) FIN-00014 University of Helsinki Finland Telephone: +358-9-191 23743
Mobile: +358 50 599 0540 Fax: +358-9-191 22 779 WWW: rni.helsinki.fi~bohOpen ↗

in article [email hidden], William Morse at
[email hidden] wrote on 24/1/04 7:51 PM:

Quoted message said:

1) Epistatic interactions between the rest of the genome on the

Quoted message said:

coefficent of selection are unknown and perhaps unknowable

Exactly what difference does this make? When a tiger is chasing Joe and Fred, and Fred is slower,
the tiger is not going to go after Joe instead because the gene that makes Joe faster has
epistatic interactions.

No but if fred is slower but carried genes that in another genetic back ground would have made him
faster than Joe or to put it another way joe carried modifiers that rendered mutations in his
"slow genes" neutral" then the "every thing works averages out in the end model" has turned out a
bad result

Quoted message said:

Quoted message said:

2) selection events are insufficiently uniform. I.e every individuals death is a unique event

The question in general is not death, it is life. Many are born, few survive. This is one of the
key principles of evolution.

I disagree here one of the key principles in evolution as change in gene frequencies is that rate at
which mutations are removed from the population. This can only be by death or reverse mutation which
is sufficiently rare to be ignored

Quoted message said:

Quoted message said:

3) the combination of the above prevents the normalisation of the data to quantify the effect of
a particular allele at a particular locus.

Quoted message said:

I would have though fisher and others may have dealt with these issues any one know of any
investigations in to this area

Others have responded to this point.

But perhaps evolution is better thought of as changes in phenotype. We typically say that
horseshoe crabs are "unchanged" after millions of years, even though their gene frequencies today
are almost certainly significantly different from their gene frequencies millions of years ago.

I couldn't agree more. But changes in phenotype are so often taken as changes in a gene that we
confuse the issue.

Returning to Joe and Fred. Fred meets a sticky end. This we know, what was the cause of death? He
was eaten by a tiger. Was it a fast tiger? If Fred had met the tigers slower mate ginger would Fred
be here today. Does that fact that Fred was only 4 years old and Joe was in his early twenties and
if compared age for age Fred would have been much faster. Or perhaps Fred was actually faster but
was not bright enough to realise he was being chased by a tiger or perhaps was running in the wrong
direction. There are so many possible confounding factors. What we do know is that the experiment
will never be repeated again. Never again will two individuals with exactly those two genotyoes at
that age will meet exactly the same tiger in exactly the same situation. So we must be very careful
what conclusions we extract from such data.

Quoted message said:

Yours,

Bill Morse

"phillip smith" <[email hidden]> wrote in message
"]news:[email hidden]...

Quoted message said:

Quoted message said:

But perhaps evolution is better thought of as changes in phenotype. We typically say that
horseshoe crabs are "unchanged" after millions of years, even though their gene frequencies
today are almost certainly significantly different from their gene frequencies millions of
years ago.

I couldn't agree more. But changes in phenotype are so often taken as changes in a gene that we
confuse the issue.

That is why I frequently use the bastardized word 'genophenotype'.

Equally, 'eitheroring' is not always a good strategy if one wants to understand what is going on.

A careful tolerance principled use of suitably smeared-out language tools - tools for covering,
scooping-up, and grasping otherwise too spread-out and slippery aspects of Reality - is sometiomes
the way to go.

Furthermore, when in comes to putting a philosophical finger on aspects of reality that are both by
nature and definition so 'foul' [i.e. so personally disturbing by associative linkage to conditioned-
in states of memory of the kind covered by expressions/terms/acronyms such as "cathected neurons"
(Freud), "primal pain" or "Pain" (Janov), "engrams" (Hubbard), or CURSES (my concEPT)] -- given our
phylogenetic, immediate intergenerational epigenetic, and personal life-time 'programming' -- that
we tend to "take flight"$ from them, then concepts built from a recipe of for fuzzsilly logic may
provide helpful relief.
%-]

I used the expression "take flight" as a metaphor for that we very commonly have come to rely on our
evolved capacity to reflexively subconsciously repress and in realtime pre-consciously 'select' our
focuses of [psychomotor] "actention"# onto - or subconsciously 'chose' to pay actention to - just
about anything *other* than aspects of reality whose unconnected internal 'neural circuit
representations' would, were they allowed to freely interact and become dynamically connected in a
way that would close these typically *kept-separate* circuits ("gated" synapses/dynamic neural
information) into constituting a profound awareness - an awareness likely to be preceded or
accompanied by deeply cognitively and emotionally disturbing (deeply distressful) emotions.

However, on the other hand, such changes of neural connexions may have a profoundly healing effect -
but *only* in cases/situations where such "neural openess" is optimal - i.e. not too open and not to
closed [i.e., an optimum state of openess relative to the 'neurodynamic pressure' exerted by the
chronically hypersensitized 'pain transmitting' neurons that are the crux of CURSES] and with a
sufficient quality and quantity of pre-existing "actention modules" available to form the new, thus
self-regulatory, functional circuits.

$ Just to make my point more clear: I meant "take flight" in philosophical and intellectual ways
*especially* - "especially", since other variants of 'AEVASIVE action' do exist.

#
Actention is an amalgam of "attention"+"action" - i.e. action(s) in the sense of both overt
behaviour and any more or less complex coordinated 'muscular activity'.

P

phillip smith <[email hidden]> wrote in
:"]news:[email hidden]:

Quoted message said:

in article [email hidden], William Morse at [email hidden] wrote on
24/1/04 7:51 PM:

Quoted message said:

Quoted message said:

But perhaps evolution is better thought of as changes in phenotype. We typically say that
horseshoe crabs are "unchanged" after millions of years, even though their gene frequencies today
are almost certainly significantly different from their gene frequencies millions of years ago.

Quoted message said:

I couldn't agree more. But changes in phenotype are so often taken as changes in a gene that we
confuse the issue.

Quoted message said:

Returning to Joe and Fred. Fred meets a sticky end. This we know, what was the cause of death? He
was eaten by a tiger. Was it a fast tiger? If Fred had met the tigers slower mate ginger would
Fred be here today. Does that fact that Fred was only 4 years old and Joe was in his early
twenties and if compared age for age Fred would have been much faster. Or perhaps Fred was
actually faster but was not bright enough to realise he was being chased by a tiger or perhaps was
running in the wrong direction. There are so many possible confounding factors. What we do know is
that the experiment will never be repeated again. Never again will two individuals with exactly
those two genotyoes at that age will meet exactly the same tiger in exactly the same situation. So
we must be very careful what conclusions we extract from such data.

But we do get non-contingent results. And in evolution we have had experiments repeated. On three
continents we have had mammals evolve in the presence of the potential food resource of anthills.
The result is spiny anteaters, pangolins, and south american anteaters. They are totally unrelated,
but share numerous features in common. Now they are not identical, because evolution _is_
contingent. But they are similar, because evolution is also statistical - even with the confounding
factors there is enough equivalent comparison (Fred is 22, Joe is 20, both turned as the tiger
started its charge, often enough that differential speed alone made a difference in survival) to
give evolution some typical directions. These directions represent possible niches - large size and
low reproductive rate, small size and high reproductive rate: speed, armor: camouflage, toxicity and
bright colors: etc. So we cannot talk about an overall direction for evolution, but we can make
predictions despite contingency.

Yours,

Bill Morse

"John Edser" <[email hidden]> wrote in message news:<[email hidden]>...

Quoted message said:

Hamilton's supposed independent gene level of selection does not exist in nature because non
additive gene fitness epistasis prohibits it. Non additive gene fitnesses must be selected
together. This fact denies any selective independence for a genomic gene.

John,

Imagine a gene (or more precisely, an allele) for "green beard altruism" - (GBA). This allele has
two effects - it causes the bearer to have a green beard, and it causes the bearer to be
particularly nice to people with green beards - at some cost to personal fitness if the niceness is
not reciprocated. The above quote makes me think that that you are dubious that the GBA gene will
proliferate. Your reasoning can be dramatized as follows: All of the other genes in the organism, at
all of the other loci, will "realize" that GBA is not a "team player". It is promoting "GBA-fitness"
rather than promoting organism fitness as it should. Therefore, they will conspire to subvert GBA
from accomplishing its selfish goal. Tactics might include producing a green beard without the
altruism, preventing the altruism from happening, and perhaps even segregation distortion to prevent
GBA from reproducing. The other genes will have no difficulty preventing GBA from succeeding - they
outnumber him.

Does this dramatization capture your reasoning properly? If it does, then I should say that I fully
agree with you. GBA will not succeed in its nefarious plot, at least not in the long term.

But now consider a gene for altruism to close relatives, or a gene for altruism to close neighbors
where the population structure makes it likely that a neighbor is also a relative. For specifity,
if necessary, consider the AMFB gene described on a different thread. What do the other genes in
the organism think about AMFB? I would claim that they are perfectly happy to let AMFB do its
thing. Although AMFB's "goal" is to benefit copies of itself in other organisms, the way it goes
about it is to benefit organisms (relatives) that are just as likely to contain copies of AMFB's
teammates as they are to contain copies of AMFB itself. So, in this case, which is the case
considered by Hamilton in his 1964 paper, non additive gene epitaxis should not be expected to work
against AMFB. Comments?

I do tend to agree with you that truly "selfish" genes - ones that decrease the inclusive fitness of
their hosts - will ultimately fail due to epitaxis. However, "clannish" genes, which promote the gene-
level interests of the entire organism's genome, without favoritism, may, as Hamilton suggests,
compete and win against "isolationist" genes that promote the host organisms narrow interests
without concern for the replicas of the host's genes that live in the host organism's relatives. Or
at least, if Hamilton's rule is wrong, I don't see that epitaxis has anything to do with it.

Jim

phillip smith <[email hidden]> wrote in message news:<[email hidden]>...

Quoted message said:

Evolution has been described as changes in gene frequencies. I have myself espoused this point of
view. I am beginning to have some doubts. The gene centered view of selection with coefficients
for each allele where the differences of genomic back ground are supposed to average out and the
frequency of the gene changes because of the mean selective advantage it confers to individuals
bearing that allele. Please ignore genetic drift for the purpose of this argument. Does any one
know of any theoretical discussion of this assertion. I can think of three confounding factors for
such as argument.
1) Epistatic interactions between the rest of the genome on the coefficent of selection are
unknown and perhaps unknowable
2) selection events are insufficiently uniform. I.e every individuals death is a unique event
3) the combination of the above prevents the normalisation of the data to quantify the effect of a
particular allele at a particular locus.

I'm curious whether you are simply doubtful that we can, in practice, test the theory that we
have, or whether you suspect that the theory is giving us a too-narrow picture of the richer-than-we-
know reality.

If you think we need a richer theory, then it appears from the discussion so far that we ought to
redefine evolution as a change in both gene frequencies and linkage coefficients. After all, we know
that selectively beneficial changes in linkage are taking place - it is glaringly obvious that this
happens in bacteria at least. In addition, I suspect that linkage changes play an important role in
speciation.

Of course, it is possible to think of each linkage as another "gene" - but these artificial linkage
"genes" are a bit odd in that the homozygotes have much higher fitnesses than the heterozygotes.
They have weird epitaxis properties too. So, I'm not sure that treating linkages as ordinary genes
is the best way to go.

I am an amateur - so I may be talking nonsense here. Is there already an acceptable treatment of the
natural selection of linkage? Can anyone provide references?

Does anything else change selectively in evolution besides gene frequencies and linkage
coefficients? I don't know.

phillip smith <[email hidden]> wrote or quoted:

Quoted message said:

Evolution has been described as changes in gene frequencies. I have myself espoused this point of
view. I am beginning to have some doubts. The gene centered view of selection with coefficients
for each allele where the differences of genomic back ground are supposed to average out and the
frequency of the gene changes because of the mean selective advantage it confers to individuals
bearing that allele. Please ignore genetic drift for the purpose of this argument. Does any one
know of any theoretical discussion of this assertion. I can think of three confounding factors for
such as argument.
1) Epistatic interactions between the rest of the genome on the coefficent of selection are
unknown and perhaps unknowable
2) selection events are insufficiently uniform. I.e every individuals death is a unique event
3) the combination of the above prevents the normalisation of the data to quantify the effect of a
particular allele at a particular locus.

Definitions of evolution involving gene frequenquencies *are* narrow.

Is it possible to do better?

Evolution /could/ be defined as being:

``Changes to the heritable elements of a population''
--
__________
|im |yler timtyler.orgtimtyler.orgOpen ↗ [email hidden] Remove lock to reply.

TT:- Definitions of evolution involving gene frequenquencies *are* narrow.

Is it possible to do better?

Evolution /could/ be defined as being:

``Changes to the heritable elements of a population''

JE:- For the above definition to represent a testable theory of evolution and not just a loose non
testable view of evolution it has to be amended to read:

" NON RANDOM changes to the heritable elements of a population "

Best Wishes,

John Edser Independent Researcher

PO Box 266 Church Pt NSW 2105 Australia

[email hidden]

in article [email hidden], Tim Tyler at [email hidden]
wrote on 3/2/04 5:55 AM:

Quoted message said:

Definitions of evolution involving gene frequenquencies *are* narrow.

Is it possible to do better?

Evolution /could/ be defined as being:

``Changes to the heritable elements of a population''
--

What is wrong with simply "change over time"

IMHO simple is better

in article [email hidden], Jim Menegay at
[email hidden] wrote on 2/2/04 8:03 PM:

Quoted message said:

I'm curious whether you are simply doubtful that we can, in practice, test the theory that we
have, or whether you suspect that the theory is giving us a too-narrow picture of the richer-than-we-
know reality.

Both. But more so the former. If the theory is indeed faulty the the later must be true. I have no
particular axe to grind but have a nagging doubt about the model

Quoted message said:

If you think we need a richer theory, then it appears from the discussion so far that we ought to
redefine evolution as a change in both gene frequencies and linkage coefficients. After all, we
know that selectively beneficial changes in linkage are taking place - it is glaringly obvious
that this happens in bacteria at least. In addition, I suspect that linkage changes play an
important role in speciation.

The problem is a technical one, science tends to look in places in can see best. To resolve these
issues we need to do population genomics. This is not practical at the moment. We can look at the
large chromosomal scale where we see genome rearrangements are fairly common. Or we can look at
small segments of DNA in detail. But we need to look at whole genomes in detail

Quoted message said:

Of course, it is possible to think of each linkage as another "gene" - but these artificial
linkage "genes" are a bit odd in that the homozygotes have much higher fitnesses than the
heterozygotes. They have weird epitaxis properties too. So, I'm not sure that treating linkages as
ordinary genes is the best way to go.

Large changes are dismised because genes are seen as bead on a string. In this model, what matters
are the properties of the beads not their position on the string. The genes as beads on a string
model has been under attack for a long time and I think is becoming untenable as predictive tool.
But what other tools do we have? Hamming distance has a certain appeal this is a measure of the
differences in two sequences. If you are unfamiliar with the concept

Tiny Tine Pine Mine Mike Nice Neck Beck Beak Beat Boat

This is the output(slightly edited) from some software I wrote to find the chain of words to link
two four letter words by changing one letter at a time. That is a change of a hamming distance of
one, at each step. The over all change is 4 The problem with hamming distance is it does not measure
transpositions, duplications, deletions and inversions well(tricks used again and again in
evolution) For example Fade and edaf (not a real word) would have a hamming distance of 4 but we can
see its a simple inversion. Like wise with an insertions etc Boat iboat the frame shift gives gives
us problems We need a new concept to take these things into account

Quoted message said:

I am an amateur - so I may be talking nonsense here. Is there already an acceptable treatment of
the natural selection of linkage? Can anyone provide references?

Does anything else change selectively in evolution besides gene frequencies and linkage
coefficients? I don't know.

This depends on your definition of evolution. I am trying to establish a good definition

in article [email hidden], Anon. at
[email hidden] wrote on 24/1/04 7:51 PM:

Quoted message said:

But you need linkage disequilibrium (LD) as well as epistasis to change the effects of a single
gene. LD is reduced by recombination, so you'd need the interacting genes to be close to each
other, or for the epistasis to be large, for the affect to be appreciable.

This is true of the epistatic interactions are limited to a single allele of a single gene at a
limited stretch of the genome. For this to be true the rest of the genome must be considered static.
I doubt this ever occurs or occurs out side of the lab with sufficient frequency to matter. To
illustrate my point take the rest of the genome save the gene we are interested in and cinsider it
all as being one gene. How many alleles are there and what is the epistatic effects on that on the
coefficent of selection of the gene we are examining?
--

Phillip Smith phills@(buggger).co.nz replace bugger with ihug applied-evolution.co.nzapplied-evolution.co.nzOpen ↗

"he who is smeared with blubber has the kindest heart" -- a Greenland Eskimo adage

Quoted message said:

Even if you do have evolution due to epistasis, this will probably entail changes in gene
frequencies anyway. It's just that you have to consider more than one gene to understand the
dynamics correctly.

In the mid 80's I was told by someone that there are no such thing as genes. At the time I thought
it was a great joke. Now I'm not so sure. If there are no genes then gene frequencies are irrelevant