Phylogeny Flowers

Today we are going to apply what we have learned about cladistics to some fictitious plant species. Below are pictures of the flowers of these plants. We have selected an appropriate outgroup (OG) to analyze our ingroup which we are calling "phylogeny flowers."

The first task is to determine which characters will be useful in our cladistic analysis. Obviously all the flowers have petals, so presence of petals would not be a useful or helpful choice. That characteristic is plesiomorphic to all of these species, not just the ingroup.

So we look over the flowers to determine how the ingroup species differ from the outgroup. We set up a data matrix with the characteristics we want to use and the taxa being observed.

Character:OGABCDE
1. Petal Color:
2. Petal Number:
3. Petal Apex:
4. Petal Length:
5. Sepals:

Now we need to determine which state of each character is apomorphic and which state is plesiomorphic...in other words we need to polarize the character states. Obviously the apomorphic state becomes altered from the outgroup in at least some members of the ingroup. So, we notice that the petals are white in the outgroup and some of the ingroup this must be the plesiomorphic color, but some of the ingroup (C, D, and E) have gray petals. So gray must be the apomorphic state. Here is the result of our polarization decisions:

Character: Apomorphic StateOGABCDE
1. Petal Color: Gray
2. Petal Number: Six
3. Petal Apex: Pointed
4. Petal Length: Long
5. Sepals: Present

The next step is to populate our data matrix with the character states in the flowers of the various taxa. Below are the flowers and the populated data matrix. Look this over and be sure you understand why some cells in the matrix have a 0 and other cells have a 1.

Character: Apomorphic StateOGABCDE
1. Petal Color: Gray000111
2. Petal Number: Six011111
3. Petal Apex: Pointed000010
4. Petal Length: Long001111
5. Sepals: Present001100

Now we examine the matrix and notice that it is possible to tally the degree to which the apomorphies are synapomorphies. This analysis is done by simply adding up the digits in each row of the table:

Character: Apomorphic StateOGABCDETotal
1. Petal Color: Gray0001113
2. Petal Number: Six0111115
3. Petal Apex: Pointed0000101
4. Petal Length: Long0011114
5. Sepals: Present0011002

As you can see above, having six petals is the most synapomorphic character... every member of the ingroup has it. This then must be a plesiomorphy for the ingroup...it must have evolved and been present in the common ancestor of the phylogeny flowers.

At the opposite extreme we have pointed-petal apex. This character state evolved recently and is found in only one taxon (D); so it is an autapomorphy.

Our next task is to use the data matrix to construct a possible cladogram. In general you start with the most synapomorphic characters and work toward the autapomorphic characters. So we start to map the evolutionary pathway by using petal number to split the outgroup from the ingroup.

____________OG
 |
 = 2 Six Petals
 |
 |___________ABCDE

OK, so far so good. Now we go to the next-most-synapomorphic character. In our data matrix that would be Long Petals, shared by four taxa. This characteristic divides our ingroup by splitting off A from the rest of the ingroup.

____________OG
 |
 = 2 Six Petals
 |
 |___________A
 |
 = 4 Long Petals
 |
 |___________BCDE

We continue by now looking at Gray Petals, shared by three taxa...this one splits B from the remaining group:

____________OG
 |
 = 2 Six Petals
 |
 |___________A
 |
 = 4 Long Petals
 |
 |___________B
 |
 = 1 Gray Petals
 |
 |___________CDE

Our next-most synapomorphic character, Sepals Present, is shared by B and C. But we have already separated B from C, so this is a problem. There is no easy way to resolve this. So we shall skip it and come back later to resolve it.

So we will go to our autapomorphy, Pointed Petals, which is found only in D. This splits off C and E from D...

____________OG
 |
 = 2 Six Petals
 |
 |___________A
 |
 = 4 Long Petals
 |
 |___________B
 |
 = 1 Gray Petals
 |
 |___________CE
 |
 = 3 Pointed Petals
 |
 |___________D

Obviously we are unable to resolve C and E until we deal with the Sepals Present issue. Clearly one explanation for this character state appearing in two different parts of the phylogeny is multiple evolutions...

____________OG
 |
 = 2 Six Petals
 |
 |___________A
 |
 = 4 Long Petals
 |
 |      5 Sepals Present*
 |______//___B
 |
 = 1 Gray Petals
 |
 |      5 Sepals Present*
 |______//___C
 |
 |___________E
 |
 = 3 Pointed Petals
 |
 |___________D

This explanation certainly will work and could even be the path of evolution, but we need to think about parsimony. Is this the simplest explanation that we can find? We have a tree with six transition steps which means there is just one homoplasy (Sepals Present*). But how easy is it to evolve sepals? How likely is this character to have evolved twice? Sepals have veins, mesophyll epidermis, guard cells, etc. Forming a whole organ is probably not very easy, and doing it twice seems very unlikely. Is it possible to explain our taxa by allowing one evolution and one reversal instead? If the development of a whole organ is really a cascade of events, it may only take one defective enzyme early in the cascade to block the whole process. In other words a reversal might be easier to explain than a second complicated evolution.

Clearly we might consider evolving sepals before B and then reversing it after C...

____________OG
 |
 = 2 Six Petals
 |
 |___________A
 |
 = 4 Long Petals
 |
 = 5 Sepals Present*
 |
 |___________B
 |
 = 1 Gray Petals
 |
 |___________C
 |
 = 5 Sepals PresentR*
 |
 |___________E
 |
 = 3 Pointed Petals
 |
 |___________D

This tree is more parsimonious...it too has six steps and one homoplasy (*) involved...but biologically one evolution and one reversal is more parsimonious than two evolutions of a complex appendage. So while a computer that designs trees might show this as equally parsimonious, a biologist knows that this is a simpler tree that has a greater probability of being the way evolution works. Remember evolution is based on very rare (1 in 1 million) mutation, so it is far more likely that a simple explanation is true than one that involves many steps happening in parallel. So it takes a biologist to resolve the homoplasies and select the most biologically parsimonious tree.

Of course, we can ask whether we have done this yet or not. Is there another tree of equal parsimony in terms of steps and homoplasies, but which is even more likely to happen biologically? Well, pigmentation is sometimes based upon the presence of a single enzyme needed to convert a colorless substrate into a colored product. Perhaps you remember how polyphenoloxidase converted colorless catechol into a red product in either Bio 115 or Bio 221. So could we construct a tree that is no more than six steps depending upon Gray Petals rather than Sepals Present?

Character: Apomorphic StateOGABCDETotal
1. Petal Color: Gray0001113
2. Petal Number: Six0111115
3. Petal Apex: Pointed0000101
4. Petal Length: Long0011114
5. Sepals: Present0011002

To follow this concept through, we try to construct a tree that uses the other characters and leaves Gray Petals to the last. So we would map Six Petals, Long Petals, Sepals Present, and Pointed Petals in that order and then figure out the homoplasy that will be presented in Gray Petals.

____________OG
 |
 = 2 Six Petals
 |
 |___________ABCDE

____________OG
 |
 = 2 Six Petals
 |
 |___________A
 |
 = 4 Long Petals
 |
 |___________BCDE

____________OG
 |
 = 2 Six Petals
 |
 |___________A
 |
 = 4 Long Petals
 |
 |___________DE
 |
 = 5 Sepals Present
 |
 |___________BC

____________OG
 |
 = 2 Six Petals
 |
 |___________A
 |
 = 4 Long Petals
 |
 |___________E
 |
 |      3 Pointed Petals
 |_______//__D
 |
 = 5 Sepals Present
 |
 |___________BC

So the last tree above just needs to be resolved for Gray Petals (found in C, D, and E). Of course we could explain this with three evolutions of the enzyme for gray pigment:

____________OG
 |
 = 2 Six Petals
 |
 |___________A
 |
 = 4 Long Petals
 |
 |      1 Gray Petals*
 |______//___E
 |
 |  1 Gray*  3 Pointed Petals
 |__//____//_D
 |
 = 5 Sepals Present
 |
 |      1 Gray Petals*
 |______//___C
 |
 |___________B

But of course this too is not very parsimonious even though it may involve just one enzyme. Evolving it three times is just unlikely. It is taking us seven steps to resolve five characters...worse than the previous attempt with Sepals Present. But could we use a reversal of pigmentation to make this tree even more appealing (biologically parsimonious)?

____________OG
 |
 = 2 Six Petals
 |
 |___________A
 |
 = 4 Long Petals
 |
 = 1 Gray Petals*
 |
 |___________E
 |
 |      3 Pointed Petals
 |________//_D
 |
 = 5 Sepals Present
 |
 |___________C
 |
 = 1 Gray Petals R*
 |
 |___________B

Now that's what I'm talking about! Here is a tree of six steps, it involves just one homoplasy, and that one is a trait that could be caused by a single gene that evolved once and then was mutated to become defective once. This final tree is the most biologically parsimonious of all! It is worthy of being presented in a more formal way:


While we might congratulate ourselves on finding an ideal tree, please understand that, just like with statistics, there is a chance that evolution did not follow the MOST parsimonious path...however small that chance may be. So we have to remember that we could be wrong here!

On the other hand, how far wrong could we be? Let's take a look at another explanation for our data matrix.

Character: Apomorphic StateOGABCDETotal
1. Petal Color: Gray0001113
2. Petal Number: Six0111115
3. Petal Apex: Pointed0000101
4. Petal Length: Long0011114
5. Sepals: Present0011002

Could this be a legitimate alternative explanation?

_____________________________OG

      2
___//_________________________A

      2           4           5
___//_____//_____//_____________B

      1           2           4           5
___//_____//_____//_____//______C

      1           2           3           4
___//_____//_____//_____//______D

      1           2           4
___//_____//_____//_____________E

Well, the diagram here is showing 15 steps instead of 6...most unparsimonious. It certainly explains the traits we see in the flowers...but how likely is it that every single trait has to appear individually as an event for each species? Yes, very unlikely. And notice how at the left edge of this diagram there are no connections between the species...in other words, no relationship? Indeed this is a model that you read about all the time in the newspaper and on television. This is the creationist's model. Their idea is that there is an "intelligent designer" that creates each species individually, and that each species then produces its own kind, and only its own kind from the point of creation through eternity. All the species were created at the same point in time literally just a day or so apart, so they all start from the left edge as individual creations. So each trait has to be created de novo in each species.

The point here is that evidence and the mechanisms we observe in nature are not consistent with this last model. And, in fact argue strongly against it. Could it be that the "intelligent designer" created a system that would and did evolve over time? What would prevent the "designer" from doing that? And, as long as you take time symbolically as suggested in at least two places in scripture, why does the sequence of what is presented there square with what we find in cladistic analysis using parsimony? Hmmm....