What is Relatedness?

Discussion in 'Health and medical' started by Jim McGinn, Feb 28, 2004.

  1. Jim McGinn

    Jim McGinn Guest

    Tim Tyler <[email protected]> wrote

    > Regarding an old, recurring SBE debate...
    >
    > Here's Dawkins on the relationship between man and chimpanzees:
    >
    > ``The 98% doesn't mean we are 98% chimpanzees. And it really matters what unit you choose to make
    > your comparison."

    Implicit in Dawkins' statement here is the supposition that the "unit" can be chosen arbitrarily to
    achieve different results in the comparison. This seems, obviously, scientifically untenable to me.
    Suppose we were comparing the length of two sticks. And suppose we use inches and determine that the
    first stick is twice as long as the second stick. If we were to employ centimeters instead of inches
    we would not expect the relative comparative lengths of the sticks to change.

    This suggests that the "units" that Dawkins refers to are not really units and/or that the scale
    that he is drawing from is greatly lacking in quantitative relevance.

    This brings two questions to mind:

    What, exactly, is relatedness? More specifically, what measurable criteria is involved with
    determining relatedness?

    What, exactly, is the quantitative scale of relatedness that we should employ so that comparative
    results of any test we might perform will be consistent regardless of the units we choose?

    Jim
     
    Tags:


  2. Tim Tyler

    Tim Tyler Guest

    Jim McGinn <[email protected]> wrote or quoted:
    > Tim Tyler <[email protected]> wrote

    > > Here's Dawkins on the relationship between man and chimpanzees:
    > >
    > > ``The 98% doesn't mean we are 98% chimpanzees. And it really matters what unit you choose to
    > > make your comparison."
    >
    > Implicit in Dawkins' statement here is the supposition that the "unit" can be chosen arbitrarily
    > to achieve different results in the comparison. This seems, obviously, scientifically untenable to
    > me. Suppose we were comparing the length of two sticks. And suppose we use inches and determine
    > that the first stick is twice as long as the second stick. If we were to employ centimeters
    > instead of inches we would not expect the relative comparative lengths of the sticks to change.
    >
    > This suggests that the "units" that Dawkins refers to are not really units and/or that the scale
    > that he is drawing from is greatly lacking in quantitative relevance.

    Dawkins is talking about the length of DNA sequence being considered.

    If you compare the similarity to two books by counting the letters that are in identical places
    between them then you get one value.

    If you compare by counting the word similarities to one another, then you get another,
    *different* value.

    The "letter" relatedness will usually be much greater than the "word" relatedness.

    The 98% figure corresponds to "letter" relatedness.

    However the "letter relatedness" of two totally unrelated *random* DNA sequences is /still/ in
    excess of 12%.

    ...and consider this example:

    "SETS GUN OFF" "LETS RUN OFF"

    Here the *letter* similarity is high - but both the word similarity and the meaning
    similarity are not.

    Sometimes, letter relatedness is not a very good metric - and it should often be considered to be
    artificially inflated.
    --
    __________
    |im |yler http://timtyler.org/ [email protected] Remove lock to reply.
     
  3. Jim McGinn

    Jim McGinn Guest

    Tim Tyler <[email protected]> wrote in message news:<[email protected]>...
    > Jim McGinn <[email protected]> wrote or quoted:
    > > Tim Tyler <[email protected]> wrote
    > > > Here's Dawkins on the relationship between man and chimpanzees:
    > > >
    > > > ``The 98% doesn't mean we are 98% chimpanzees. And it really matters what unit you choose to
    > > > make your comparison."
    > >
    > > Implicit in Dawkins' statement here is the supposition that the "unit" can be chosen arbitrarily
    > > to achieve different results in the comparison. This seems, obviously, scientifically untenable
    > > to me. Suppose we were comparing the length of two sticks. And suppose we use inches and
    > > determine that the first stick is twice as long as the second stick. If we were to employ
    > > centimeters instead of inches we would not expect the relative comparative lengths of the sticks
    > > to change.
    > >
    > > This suggests that the "units" that Dawkins refers to are not really units and/or that the scale
    > > that he is drawing from is greatly lacking in quantitative relevance.
    >
    > Dawkins is talking about the length of DNA sequence being considered.

    The fact that Dawkins and yourself are unable to delineate this "length" proves my point.

    Jim
     
  4. Tim Tyler

    Tim Tyler Guest

    Jim McGinn <[email protected]> wrote or quoted:
    > Tim Tyler <[email protected]> wrote in message news:<[email protected]>...
    > > Jim McGinn <[email protected]>
    > > > Tim Tyler <[email protected]> wrote:

    > > > > Here's Dawkins on the relationship between man and chimpanzees:
    > > > >
    > > > > ``The 98% doesn't mean we are 98% chimpanzees. And it really matters what unit you choose to
    > > > > make your comparison."
    > > >
    > > > Implicit in Dawkins' statement here is the supposition that the "unit" can be chosen
    > > > arbitrarily to achieve different results in the comparison. This seems, obviously,
    > > > scientifically untenable to me. Suppose we were comparing the length of two sticks. And
    > > > suppose we use inches and determine that the first stick is twice as long as the second stick.
    > > > If we were to employ centimeters instead of inches we would not expect the relative
    > > > comparative lengths of the sticks to change.

    [...]

    > > Dawkins is talking about the length of DNA sequence being considered.
    >
    > The fact that Dawkins and yourself are unable to delineate this "length" proves my point.

    Your point /seemed/ to be that Dawkins was saying:

    ``that the "unit" can be chosen arbitrarily to achieve different results.''

    That's not what the man said - he was pointing out that comparing chunks of text of different length
    would give different values for "relatedness" - in much the same way that comparing words and
    comparing letters between two books does.

    He never said that all the resulting values were equally valid or useful - and thus that the length
    could be chosen arbitrarily.

    Indeed what he said was just the opposite - he said that: "it really matters what unit you choose" -
    since comparing stretches of genome of different lengths can give different results.

    He also said that the "letter" relatedness of "98%" did *not* mean that we were "98%" chimpanzees -
    saying that this sort of relatedness has drawbacks.
    --
    __________
    |im |yler http://timtyler.org/ [email protected] Remove lock to reply.
     
  5. Peter F.

    Peter F. Guest

    "Jim McGinn" <[email protected]> wrote in message
    news:[email protected]...
    > Tim Tyler <[email protected]> wrote in message
    news:<[email protected]>...
    > > Jim McGinn <[email protected]> wrote or quoted:
    > > > Tim Tyler <[email protected]> wrote
    > > > > Here's Dawkins on the relationship between man and chimpanzees:
    > > > >
    > > > > ``The 98% doesn't mean we are 98% chimpanzees. And it really matters what unit you choose to
    > > > > make your comparison."
    > > >
    > > > Implicit in Dawkins' statement here is the supposition that the "unit" can be chosen
    > > > arbitrarily to achieve different results in the comparison. This seems, obviously,
    > > > scientifically untenable to me. Suppose we were comparing the length of two sticks. And
    > > > suppose we use inches and determine that the first stick is twice as long as the second stick.
    > > > If we were to employ centimeters instead of inches we would not expect the relative
    > > > comparative lengths of the sticks to change.
    > > >
    > > > This suggests that the "units" that Dawkins refers to are not really units and/or that the
    > > > scale that he is drawing from is greatly lacking in quantitative relevance.
    > >
    > > Dawkins is talking about the length of DNA sequence being considered.
    >
    > The fact that Dawkins and yourself are unable to delineate this "length" proves my point.

    As far as my primate (near but not quite chimp-style) brain can overview the relevant science
    factual situation, you, Jim, should not expect to find answers to your questions any time soon.

    In fact it even appeared [when I percEPTively sniffed the wind - using not my nose but an organ that
    produces playfully oracular opinions - in regard to these questions %-] that there will NEVER arise
    any such - 'statistically straight-forward' - answers (for us to champion our superiority over
    chimpanzees with) other than already available and collectable of data by which we can determine
    that we are the champions as far as AEVASIVE adaptability/capacity is concerned.

    P
     
  6. Jim McGinn

    Jim McGinn Guest

    Tim Tyler <[email protected]> wrote

    <snip>

    > > > Dawkins is talking about the length of DNA sequence being considered.
    > >
    > > The fact that Dawkins and yourself are unable to delineate this "length" proves my point.
    >
    > Your point /seemed/ to be that Dawkins was saying:
    >
    > ``that the "unit" can be chosen arbitrarily to achieve different results.''
    >
    > That's not what the man said - he was pointing out that comparing chunks of text of different
    > length would give different values for "relatedness"

    Did he not use the word, "units?"

    >- in much the same way that comparing words
    > and comparing letters between two books does.
    >
    > He never said that all the resulting values were equally valid or useful - and thus that the
    > length could be chosen arbitrarily.

    It seems he was saying that the "unit" could be chosen arbitrarily to achieve whatever end
    one wished.

    > Indeed what he said was just the opposite - he said that: "it really matters what unit you choose"
    > - since comparing stretches of genome of different lengths can give different results.

    And you think this is a good thing?

    > He also said that the "letter" relatedness of "98%" did *not* mean that we were "98%" chimpanzees
    > - saying that this sort of relatedness has drawbacks.

    What's needed is a more rational and consistent understanding of relatedness. After this is achieved
    we can then try to narrow it down to biological relatedness.

    Jim
     
Loading...