transition bias

Mark Siddall mes at
Sun Jul 24 14:40:13 EST 1994

>Jonathan F Wendel (jfw at wrote:
>: transition/transversion biases.  I am wondering, though, whether there is
>: any information on biases WITHIN transitions.  Specifically, are there data
>: on AG versus CT transition frequencies in either plants, fungi, or animal
>: nuclear or organellar genomes?
In article <1994Jul23.204826.20033 at> brianf at (Brian Foley) writes:
>	Is this a trick question?  I always thought that an A -> G transition
>on one strand of DNA was accompanied by a T -> C transition on the other
>strand.  So A -> G must equal T -> C and visa versa.

Hang on a second...  usually when referring to transition/transversion
biases one is talking about only ONE strand (i.e., the coding strand),
and yes there are differences within transitions and within transversions

Beware though that how these are interpreted depends a lot on the 
perspective on the data.  For example, whether a transition or a transversion
has occurred and then what particular type depends on reference to
some phylogenetic tree.  If you get the tree wrong, you'll get the 
proportions wrong.  Perhaps more importantly for the purposes of 
weighting, you need a tree to be able to set the realtive weights (inverse
to their proportions) in order to get a tree - sound circular?

The standard approach in phylogenetic methodology is either the 
Jukes-Cantor one-parameter model in which transitions and transversions
are of equal weight or the Kimura two-parameter model in which 
transversiona are weighted the same and transitions are weighted the 
the same but differnt from transversions.

Here are some:  (left column = "from"; top row = "to")

          A  T  C  G
     A    -  a  a  a
     T    a  -  a  a
     C    a  a  -  a
     G    a  a  a  -

Kimura 2-parameter
          A  T  C  G
     A    -  b  b  a
     T    b  -  a  b
     C    b  a  -  b
     G    a  b  b  -

          A  T  C  G
     A    -  b  c  d
     T    a  -  c  d
     C    a  b  -  d
     G    a  b  c  -

          A      T       C      G
     A    -      (1-y)b  (y)b  (y)a
     T    (1-y)b  -      (y)a  (y)b
     C    (1-y)b (1-y)a  -     (y)b
     G    (1-y)a (1-y)b  (y)b  -

   - where y is the G+C content

There are other models.  And then there is the Dynamic Weighting of Fitch
and Ye (1992) (in Miyamoto and Cracraft Phylogenetic Analysis of DNA Sequences)
in which all types of changes receive weights inverse to their own 

Incidentally, the "Trace Changes and Stasis..." option in MacClade is
ideal for quantifying the amounts of each substitution type implied
on a particular tree.


Mark E. Siddall                "I don't mind a parasite...
mes at                    I object to a cut-rate one" 
Virginia Inst. Marine Sci.                     - Rick
Gloucester Point, VA, 23062

More information about the Mol-evol mailing list