24.02.2013 Views

Encyclopedia of Evolution.pdf - Online Reading Center

Encyclopedia of Evolution.pdf - Online Reading Center

Encyclopedia of Evolution.pdf - Online Reading Center

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

cladistics<br />

The above examples used easily visible characteristics to<br />

classify organisms into groups. The characteristics need not<br />

be visible. In fact, many cladistic analyses use DNA sequence<br />

information (see bioinformatics; DNA [evidence for<br />

evolution]). Suppose the scientist has segments <strong>of</strong> DNA<br />

from three species. If the segment <strong>of</strong> DNA is 100 nucleotides<br />

in length, this allows 100 characters to be used to test<br />

each cladogram. Nearly every phylogenetic study includes at<br />

least one analysis that uses DNA nucleotides as traits. It is<br />

no easier to use DNA than to use visible characters in cladistic<br />

analysis. The reason is that one cannot compare just any<br />

old piece <strong>of</strong> DNA <strong>of</strong> one species with any old piece <strong>of</strong> DNA<br />

in another, any more than one can make comparisons <strong>of</strong> just<br />

any visible trait. Just as visible traits must be homologous,<br />

so the DNA segments must be homologous. The investigator<br />

must compare the same gene in all <strong>of</strong> the species in the analysis.<br />

Further complications arise when one considers that some<br />

mutations exchange one nucleotide for another, while other<br />

mutations add or delete nucleotides, or reverse segments, or<br />

even move segments to different places. Usually, researchers<br />

use certain well-studied DNA sequences rather than trying to<br />

figure a new one out for themselves. In comparisons among<br />

plants, matK (the gene for an enzyme that acts upon the<br />

amino acid tyrosine) and rbcL (the gene for the large subunit<br />

<strong>of</strong> the enzyme rubisco) are frequently used.<br />

Some analyses use noncoding DNA for phylogenetic<br />

analysis. Researchers may use transposable elements (see<br />

horizontal gene transfer) because homoplasy is very<br />

unlikely to occur with them. It is very unlikely that the same<br />

transposable element would evolve twice: It would neither<br />

insert itself in the same location in the chromosomes <strong>of</strong> two<br />

different species, nor transfer to a new location without gaining<br />

or losing a few nucleotides in the process. Other analyses<br />

use short or long interspersed sequences in the noncoding<br />

DNA. The figure on page 73 uses the presence or absence <strong>of</strong><br />

20 segments <strong>of</strong> noncoding DNA to construct a cladogram <strong>of</strong><br />

mammals. Cows and deer are clustered together, because they<br />

share all <strong>of</strong> the segments for which information is available.<br />

Whales and hippos are clustered together, because they share<br />

15 <strong>of</strong> the 17 segments for which information is available.<br />

Whales and cows are not clustered closely together, because<br />

they share only nine <strong>of</strong> the 17 segments for which information<br />

is available. Camels are not closely related to cows, as<br />

they have none <strong>of</strong> the 11 DNA segments that cows possess.<br />

Choosing the best cladogram<br />

A computer can generate many possible cladograms for any<br />

set <strong>of</strong> species that are being compared. Which <strong>of</strong> these cladograms<br />

is the correct one? The correct one is the one that most<br />

closely matches what actually happened during evolutionary<br />

history. Nobody can know exactly what happened during<br />

evolutionary history; cladists must make an assumption about<br />

which cladogram is best. They <strong>of</strong>ten make the assumption<br />

<strong>of</strong> parsimony (see scientific method), that is, they choose<br />

the simplest explanation. To do this, they choose the cladogram<br />

that requires the fewest number <strong>of</strong> character changes.<br />

Consider this simplified example in the classification <strong>of</strong> dogs,<br />

cats, and lions. Since each group has two branches, one could<br />

classify them in three ways:<br />

1. [Dogs] [Cats, Lions]<br />

2. [Dogs, Cats] [Lions]<br />

3. [Dogs, Lions] [Cats]<br />

Cats and lions both have retractable claws; dogs do not.<br />

Cladograms 2 and 3 require that retractable claws should<br />

have evolved twice, separately in cats and lions. Therefore<br />

cladogram 1 is the most parsimonious. It would not be safe<br />

to draw a conclusion based upon just one characteristic (in<br />

this case, retractable claws). Cladograms based on any <strong>of</strong> a<br />

number <strong>of</strong> other characteristics would lead to the same conclusion:<br />

Cats and lions form the feline group, while dogs are<br />

part <strong>of</strong> a separate (canine) group. From this one would conclude<br />

that cats and lions shared a common ancestor that lived<br />

much more recently than the common ancestor <strong>of</strong> canines<br />

and felines.<br />

It is easy enough to compare the three cladograms that<br />

are possible with three species. However, as more species are<br />

compared, the analysis becomes exponentially more difficult.<br />

When comparing even a slightly larger number <strong>of</strong> species, the<br />

calculations become so complex that they can only be performed<br />

by a computer. As the number <strong>of</strong> species in the analysis<br />

increases, even a computer cannot sort through all <strong>of</strong> the<br />

possibilities. For 10 species, there are more than two million<br />

possible cladograms; for 50 species, there are 3 × 10 76 possible<br />

cladograms. To provide some idea <strong>of</strong> how large a number<br />

this is, consider that the universe has existed for less than<br />

10 18 seconds! The computer must randomly generate a subset<br />

<strong>of</strong> several hundred possible cladograms and make comparisons<br />

among them.<br />

Even when a cladistic analysis compares only a few hundred<br />

<strong>of</strong> the possible cladograms, there may be a large number<br />

<strong>of</strong> almost equally parsimonious ones. The computer must<br />

then generate a consensus tree that preserves the patterns that<br />

the best cladograms share in common with one another.<br />

Applying cladistics to classification<br />

A cladistic approach to classification would group organisms<br />

together on the basis <strong>of</strong> the branch points in the cladogram.<br />

A clade (whether it corresponds to a genus, a family,<br />

an order, a class, a phylum, or a kingdom; see Linnaean<br />

system) must consist <strong>of</strong> all species that share a common<br />

ancestor, and only those species. Such a clade is called monophyletic<br />

(mono- means one). Cladists will not accept a genus,<br />

family, order, class, phylum, or kingdom that includes only<br />

some <strong>of</strong> the species that have evolved from a common ancestor.<br />

Such a group is called paraphyletic (para- means beside).<br />

Cladists most vigorously reject taxonomic groups that both<br />

omit some species that share a common ancestor and include<br />

some species that do not, a group referred to as polyphyletic<br />

(poly- means many).<br />

A cladistic approach can lead to a different system <strong>of</strong><br />

classification than what is traditionally recognized, as in these<br />

examples:

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!