Use of secondary structure in alignments

Mark Siddall mes at vims.edu
Fri May 24 15:20:53 EST 1996


Diane Stothard wrote:
> 
> I was wondering what the consensus feeling is in this group for using 
> secondary
> structure as a guide for aligning rRNA gene sequences. I personally feel it is

Diane,
You, of course, already know how I feel about this.
1) using secondary structure presupposes that the models being used for 
alignment purposes are correct notwithstanding that the models are not proven 
correct for the tax(on/a) under consideration and that said thermodynamic models 
have been suggested as naively incorrect by others.  
2) given the foregoing, secondary structure can be used effectively in the 
conserved regions of the gene only - yet in these regions of the gene we really 
don't need its help.
3) as phylogeneticists we are interested in uncovering patterns not imposing 
patterns - use of secondary structure imposes patterns, I would argue 
unecessarily.

> necessary in order to insure comparison of homologous sites, but I have met
> many people that do not believe in secondary structures and therefore, align
> their sequences randomly (i.e., by base letter). I would especially appreciate

This is highly unfair and a gross mischaracterisation.  The use of software like 
Malign which is explicitly based on a parsimony optimality criterion using the 
Needleman-Wunsch dynamic algorithm is hardly RANDOM!  Nor does levelling such
a slander make it so.


Mark



More information about the Mol-evol mailing list