functional_analysis.Rmd

Functional trait analysis methods
=================================

Dummy trait data
----------------

Consider the following trait data (equivalent file [here](traits.tsv)):

|     | num1 | num2 | fac1 | fac2 | ord1 | ord2 | bin1 | bin2 |
|-----|------|------|------|------|------|------|------|------|
| sp1 | 9    | 4.5  | A    | X    | 3    | 2    | 0    | 1    |
| sp2 | 8.1  | 6    | A    | Z    | NA   | 1    | 0    | 1    |
| sp3 | NA   | 2.3  | C    | Y    | 5    | 3    | 1    | 1    |
| sp4 | 3.2  | 5.4  | B    | Z    | 1    | 7    | 0    | 0    |
| sp5 | 5.8  | 1.2  | C    | X    | 2    | 6    | NA   | 0    |
| sp6 | 3.4  | 8.5  | C    | Y    | 2    | 1    | 1    | 1    |
| sp7 | 7.5  | 2.1  | B    | X    | 3    | 2    | 1    | 0    |
| sp8 | 4.3  | 6.5  | NA   | Z    | 1    | 3    | 0    | 1    |

- 8 species (rows, sp1 .. sp8), 8 traits (columns)
- some traits are continuous numerical values (`num1`, `num2`)
- some are factors (`fac1`, `fac2`)
- some are ordinal, i.e. (ranked) integers (`ord1`, `ord2`)
- some are binary, e.g. presence/absence (`bin1`, `bin2`)
- most of them have some missing values (`NA`)

Pairwise distances in "trait space"
-----------------------------------

Let's read the trait data and compute the "Gower distance"" between the species.
`FD::gowdis` computes the Gower (1971) similarity coefficient exactly as
described by Podani (1999), then converts it to a dissimilarity coefficient
by using D = 1 - S. It integrates variable weights as described by
Legendre and Legendre (1998).

```{r gowdis}
library(FD)
traits <- read.csv2("traits.tsv", header = T, sep = "\t", row.names = 1 )
gdist <- gowdis(traits)
gdist
```

This gives us a pairwise, symmetrical (i.e. triangular) distance matrix, which
might be used, for example, to cluster species by functional similarity:

```{r cluster}
plot(nj(gdist), type = "unrooted")
```

Functional dispersion
---------------------

Now let's have a look at the functional dispersion. As we saw, `FDis` takes abundances into
account. Consider these abundance data (file [here](abundances.tsv)):

|      | sp1 | sp2 | sp3 | sp4 | sp5 | sp6 | sp7 | sp8 |
|------|-----|-----|-----|-----|-----|-----|-----|-----|
| com1 | 1   | 1   | 0   | 0   | 4   | 2   | 0   | 0   |
| com2 | 0   | 0   | 0   | 2   | 1   | 0   | 0   | 5   |
| com3 | 2   | 0   | 0   | 0   | 0   | 1   | 0   | 3   |
| com4 | 1   | 0   | 7   | 0   | 0   | 0   | 0   | 0   |
| com5 | 0   | 0   | 2   | 3   | 3   | 0   | 0   | 0   |
| com6 | 0   | 3   | 0   | 0   | 5   | 6   | 1   | 6   |
| com7 | 3   | 5   | 0   | 3   | 0   | 0   | 0   | 0   |
| com8 | 0   | 0   | 0   | 0   | 6   | 2   | 1   | 2   |
| com9 | 4   | 1   | 1   | 3   | 0   | 0   | 2   | 0   |

If we read the abundance data and compute `fdisp` we get estimates for the functional
dispersion of the various communities:

```{r fdisp}
abundances <- read.csv2("abundances.tsv", header = T, sep = "\t", row.names = 1 )
dispersion <- fdisp(gdist,as.matrix(abundances))
dispersion
```

References
----------

- **JC Gower**, 1971. A general coefficient of similarity and some of its properties. _Biometrics_ **27**:857-871
- **E Laliberté & P Legendre**, 2010. A distance-based framework for measuring functional diversity from multiple 
  traits. _Ecology_ **91299**:305.
- **P Legendre & L Legendre**, 1998. _Numerical Ecology_. 2nd English edition. Amsterdam: Elsevier.
- **J Podani**, 1999. Extending Gower's general coefficient of similarity to ordinal characters. _Taxon_ **48**:331-340.