Several questions on evolution, and mutation (rate)

Ram Samudrala ram at
Thu Sep 5 17:13:22 EST 1996

James Foster (foster at wrote:

>By the way, I'm only a biologist "in a sense".  I'm a computer scientist
>by training, but I've been doing computational biology work for a few

Ditto, though my other major was genetics.

>Why should you?  mutations which lead to changes that increase in
>frequency due to selective pressure are "beneficial", while those that
>don't aren't.  Of course, this classification can change overnight with
>the advent of a drought or an invasion of predators.

You can have cumulative mutations, that produce some effect.  I
consider each step in that process a mutation, and it's a beneficial
one, if a single mutation helps to benefit.  But this is not
necessarily the case.  Say in the case of antibiotic resistance, that
you needed three amino acid mutations for a specific protein that
could potentially perform this function (even though it can't
presently) to even begin binding the antibiotic and inhibiting it (if
you'd like to work with specific examples, the one I am most familiar
with are the beta-lactamases).

>And what's a "new" function, anyway?  Usually mutations only
>incrementally change some already-present characteristic, like
>sensativity to light, resistance to antibiotics, or length of beak.

That's a new function.

>Wouldn't resistance to an antibiotic, or ability to produce viable
>offspring, be new functions?

The evolution of resistance to antibiotic is a new function.  The
selection of it is a different story.  This is the distinction I'm
trying to make.  

>These rarely spring into being de novo.

I am not sure if we're communicating.  There is a situation where a
protein inhibits an antibiotic (even feebly) and where a protein
doesn't inhibit that antibiotic.  Between those points, evolution (not
natural selection) has occurred.  Say you have two related
antibiotics.  Currently, in a population of bacteria you see
resistance to one antibiotic, but not to the other.  In a few years,
you begin to see resistance to another antibiotic in that same
population.  Now, there are two explanations (in this context) for
what happened: one, the gene for resistance to the second antibiotic
was ALREADY PRESENT (as a variant of the first gene, either in the
population or on the genome) in the bacterial population, and was
NATURALLY SELECTED for.  Two, the gene for resistance EVOLVED from the
first gene, in some manner, and was THEN natural selected for.  These
are two different processes.

>One day, people are alive.  The next day, they die.  Why?  Because
>the bacteria in their bodies have developed the function of surviving
>massive doses of a common antibiotic.  Isn't that what you're after?

Nope.  Because I'd argue a few bacteria (a single one will do) already
had this function, and they were just selected for, and that bacteria
(which is resistant) reproduced to a point where most of the bacteria
in your body were not susceptible to antibiotics, thus killing the

>I don't understand your objection.  Being selected for changes the
>frequency of this character, and THAT is evolution. 

That's where we disagree.  To me, evolution is the evolution of the
function at a molecular level, that is cumulative mutations (it could
be one) that result in new function for a given protein.  Natural
selection is the selection (positive or negative) of that organism
with that protein.  Evolution operates at a molecular/genetic level.
Natural selection operates on the usefuless of the evolution that has
occured (at various levels).

>For example, genes which code for one protein in hemoglobin may
>silently duplicate, and only over much time will that duplication
>develop some functionality.

This is what I mean by evolution!  Suppose our haemoglobin genes
tomorrow also conferred resistance to malaria (by having a few
mutations).  Wouldn't you wonder HOW a new function arose from this
duplication (which must've been both haemoglobin initially), and how
long it took just to come up with it?

>It's pointless to insist on isolating the original duplication event,
>because it was silent.

No, not the original duplication event, but the length of time (and
how) it takes to come up with the new functionality (which has to
happen BEFORE natural selection).

>The important point is that mutation without selection is not a
>particularly interesting concept, from an evolutionary perspective.

We disagree there too, and there's quite a bit of research on this
topic.  Basically, the idea is, how long does it take for new function
to evolve, from an existing gene, before naturally selection can
operate on it.  How does it evolve?  Is it randomly (the current
belief) or is there some directed process going on?


me at  ||  ||
  When the going gets tough and the stomach acids flow. 
  The cold wind of conformity is nipping at your nose.
  Some trendy new atrocity brings you to your knees. 
  Come with us and we'll sail the seas of cheese. ---Primus

More information about the Mol-evol mailing list