Cliff Lundberg responds:
>Doesn't this approach work as well in a purely evolutionary context?
>Given the Cambrian boom and subsequent gradual evolution, a molecular
>model involving sudden generation of complexity and later gradual
>loss in different ways in different lineages seems reasonable.
>
>> Still I see, and won't be surprised with further work, that
>> there are patterns to the losses. i.e. if all the members of one Class all
>> have the identical 15bp deletion it seems improbably that all 100 or so
>> samples all lost the same 15 bps of DNA independently.
>
>Why improbable? You say the loss is due to functional constraints;
>aren't all 100 or so suffering the same functional constraints?
Yes, I see how my examples might lead to that possible conclusion. Here I
was thinking more of very small insertion and deletions (although I used 15
as an example) which are shared by many many species. These small
insertions are likely to occur in loop regions that have relaxed functional
constraints and so the presence of deletions in themselves is not highly
unusual and in fact has been seen in presumably non-related taxa. The
problem is that many of these deletions occur with very specific boundaries
where a common boundary is not required (i.e. there are many position free
to vary) and the same samples which may have several deletions in common
will also have several insertions in common.
To give a more specific example in the 19S SSU rDNA of plant mitochondria
there are two separate insertions (relative to all other plants) that are
unique to ferns (a 25bp insertion and a 36bp insertion). These insertions
in all lineages of ferns (including lepto- and eu-sporangiate ferns) are
found in the identical position of the 19S secondary structure. At the
same time all of the ferns have a smaller V7 and V1 domain due to the
deletion of several stretches of sequence relative to all other plants.
functionally the regions in which this variability is occurring are
unconstrained relative to the rest of the molecule. If the insertions all
represent independent gains of sequence it seems odd that they have gained
nearly the same sequence at the same position. Also Equisetum (a presumed
extant relative of the ancestors of ferns) has small insertions at these
same positions. In this example we could again posit that all of the ferns
were created with the insertions and deletions (relative to other plants)
as they are but there presence seems to make little differnce. The ITS
(Inter Transcribed Spacer) DNA between the SSU and LSU of bacteria and
plants is untranslated sequence (spacer) and it also exhibits many small
deletions and insertions that also reflect a possible ancestral
relationship between species. Here it is not thought that functional
constraints have any significant impact on the direction of evolution in
these sequences.
Its not just the presence of insertions and deletions that is most
puzzling. I would agree that one could always say that they were either
created with or without particular sequences or that they have
independently lost sequences for a particular reason. What is more
perplexing is situations that appear to require more than one step to go
from one to the other. This is what I tried to model in my hypothetical
example.
Inversions on top of insertions and further inversions are more difficult
to deal with. Take for example the follwing genes in the following order:
_________________________ inversion
sample 1) A--B--C--D--E--F--G--H--I--J--K--L--M--N--O--P
________________ inversion
sample 2) A--B--C--D--M--L--K--J--I--H--G--F--E--N--O--P
| |
sample 3) A--B--C--D--E--F--G--H--N--E--F--G--H--I--O--P
______ inversion
sample 4) A--B--C--D--E--F--G--H--N--E--F--G--H--P--O--I
If we saw these gene orders in three bacterial plasmids I don't think
anyone would have much of a problem reconstructing a recombinant history in
which two sequential inversions took place from samples 1-4 or 4-1, or if 2
was the ancestor we could have two inversions at different times (cell
divisions) resulting in 3 and 1 and later 3-4. Yet when I look at four
chlorplast genomes in which similar inversions can take place I am supposed
to look at the same inversions as representing three separately created
gene orders. At the very least there is the appearance of a history of
multiple nested inversions.
ramble mode off,
Joel duff