chloroplast and other genes / total evidence

[pierre deleporte]

At 00:17 03/04/2004 -1000, Richard Pyle wrote :
>I also predicted that our technological ability to obtain whole genomes
>cheaply and quickly will arrive somewhat before our ability to reliably
>extract the phylogenetically useful information (e.g., we already know the
>whole human genome, yet we have barely begun to understand how to interpret
>the information for constructing a human being from it).  I suspect that
>when we do understand how to extract the information, we will come to
>realize that the WHOLE genome is not really needed to be able to infer a
>phylogeny that approaches the "confidence asymptote" (i.e., the greatest
>confidence in evolutionary affinities that can possibly be achieved).  So,
>in a sense, I agree with the inference intended by the comment above: that
>simply having access to complete genomes would not provide all us with all
>of the answers ('though I'm not sure I've ever met anyone who actually
>believed that it would -- even 10 years ago).
>
>However, I guess my underlying point is that our current efforts to
>reconstruct phylogenies, and our arguments about what approaches are most
>"reliable" etc., will probably seem rather quaint a few decades hence.
>That's not to say that they're a waste of time (we have to figure it out
>somehow). But the level of passion that sometimes lies behind these debates
>seems a bit out of place and unwarranted, in that context.

In the same general line :
The popularized principle of using "total evidence" for phylogeny inference 
means (at best) using total *relevant knowledge*, not indiscriminately 
constructing the hugest possible data matrix.
Garbage in, garbage out... but how to identify garbage, or potentially 
misleading noise, from presumably informative data? In direct connection 
with this is the problem of how to interpret different lines of potentially 
informative evidence (= using the "relevant evolutionary models" for 
different kinds of data).
The practice of using "all characters we can grasp" with little confidence 
in the evolutionary model we implement is but a sign of immaturity of the 
field of phylogeny inference. Progress must come from improved 
understanding of evolutionary processes (genomics, genetics of individual 
development, population genetics, and all what can be inferred through 
comparative biology in the light of progresses in these and other fields of 
biology).
Viewed this way, rushing for extensive complete genomes sequencing (or 
exhaustive morpho-anatomical descriptions either) could effectively appear 
like putting the cart before the horse. Sequencing some complete genomes 
seems interesting only in that it provides a larger spectrum of potential 
evidence in which to search after some unsuspected kinds of optimally 
informative data.
Indiscriminate analyis of all we can scratch is likely an infantile 
sickness of phylogeny inference. People who try to freeze the practice of 
analysing all available data indiscriminately as being "the panacea" tend 
to obscure the need for improved understanding of evolutionary processes, 
hence improving relevant knowledge and making optimal use of it, which is 
the correct acception of the "total evidence" principle.

Possible reading dealing with this question:
Lecointre and Deleporte - *Total evidence requires exclusion of 
phylogenetically misleading data* - Zoologica Scripta. In press.

best,
Pierre


Pierre Deleporte
CNRS UMR 6552 - Station Biologique de Paimpont
F-35380 Paimpont   FRANCE
Téléphone : 02 99 61 81 66
Télécopie : 02 99 61 81 88