Constructing pylogenies

The Character Matrix                                               

Characters
        In essence, you would want to look at as many characters as possible; you would certainly like to look at many more character states than there are species you are trying to investigate.  In this example there are eleven characters that have been investigated.  The best characters are typically discrete traits, as continuous traits can vary and present the investigator with a continuum (which would be divided into separate states arbitrarily).  NEVER list states such as large/small, light/dark, long/short, etc.  In order to polarize the character states, you need an outgroup -- a species/group of species that is considered relatively closely related that allows you to get an idea of the "ancestral" state of the characters your are investigating, which in turn allows you to polarize the traits.

Character States

0 :  This indicates the presumptive ancestral state
1 :  This is the state that is presumed to be derived from the ancestral trait (a potential apomorphy)
1a/1b :  These are different derived character states from the ancestral character (independent of each other).
2 :  This is the state that is presumed to be derived from state 1 (though be careful about presumptions).
2a/2b:  These are different derived character states from character state 1 (independent of each other).

    and so on . . . .

So, understand that if species has character state "2" for a particular trait, then ancestrally that species had to have "passed through" character state "1" to get to "2". Therefore, character state "1" would unite this species with any other species that still have character state "1".

Having said that, then the following character states unite the following species:

1(1) :  A, B, C, F, G

2(1) :  D, E

3(1) :  E, H

4(1) :  A - G
        4(2) :  A, C, F, G

5(1) :  A - G
        5(2) :  D, E

6(1) :  D

7(1) :  A - E, G, H        (though as you will find out shortly, species F actually has a reversal of this trait)
        7(2) :  C

8(1a) :  D, E
8(1b) :  A, C, G

9(1) :  A - H (does not include the outgroup [OG])

10(1) :  A - G
        10(2a) :  B
        10(2b) :  E

11 :  no relevant information available from this trait

In class construction of a phylogeny.

        Analyze which species share which character states for EACH trait.  The analysis should allow you to make calculations as to who is most closely related to whom and construct the tree that is most parsimonius for the group of species you are investigating.  We will do this IN CLASS, with the above character matrix.

Some observations: 
1.  For the ingroup of 8 species, character 9 (and 7) roots the ingroup, and allow for polarization of the character states.
2.  H seems to be very different from the other species, as A - G are united by three character states [4(1), 5 (1), and 10(1)] that exclude H.  That would seem to indicate that H branches off basally on the tree, as indicated.
3. D and E seem to be very closely related and united by 2(1), 5(2), and 8(1a).
4.  Similarly, the other species (A - C and F - G) are also united by one character, 1(1) -- so the two main branches on the tree are [A - C, F - G] and [D & E], as indicated.
5.  A, C, F, and G are united by one character 4(2) (excluding B), resulting in that further branch on the indicated tree, and, of those, A, C, and G are further united by one character 8(1b) (excluding F), which results in the final branch as indicated.

    Any unresolved polytomies?

    Any reversals? 

    Any independent derivations of the same trait?

    Any species not supported by an autapomorphy (from the data set)?

    How many genera supported by the data set (you will need the phylogeny to tell)?

A couple more character matrices to practice with: