AFLPMax needs a reference tree to work. The tree has to be in Newick format (the same as used by PHYLIP), it has to be unrooted and
it has to be written in only one line, without spaces. Here are two examples:
Taxa names cannot start with a number but any alphanumeric character is admitted in any other position of the name.
The first and only step needed to run the program is to load a tree. You can paste it directly into the INPUT TREE tab or you can load it from the
File menu, option Open, also with the shortcut Ctrl+O. Once you have loaded a tree the Run button is activated and you can run an analysis
using the default options or you can change them in the settings button and menu.
The reference tree defined by the user (see Input file section) is used to obtain DNA sequences with the program Seg-Gen (Rambaut & Grassly, 1997).
Seq-Gen evolves sequences along a specific tree under a given evolution model (chosen by the user).
A computer program (aflp_seqgen) written in C is then used to simulate the AFLP technique. Each sequence corresponding to each taxon is cut separately
with enzymes EcoRI and MseI, which are the typical enzymes used in AFLP studies. The program searches for all the restriction sites and returns all
fragments that would result from the digestion. The output is the list of fragments sorted by their length, as it would appear in a real experiment.
The in silico AFLP profiles are generated using the 4096 possible combination of three selective nucleotides for each enzyme. Another program (aflp_phylogeny),
also written in C, is used to combine the information of all individuals and to construct the 1/0
presence/absence matrices. The program selects randomly the different number of bands used to reconstruct the phylogenies, starting from 100
and increasing by 100 each time (each set of 100 bands would correspond to a different primer combination) until a maximum of 1000 bands.
Those numbers are chosen because 100 bands per AFLP profile are recommended for experimental data sets.
AFLP-based phylogenies are estimated with programs from the PHYLIP package (Felsenstein, 2005), using two of the most widely used methods in AFLP
data sets: minimum evolution (ME) and maximum parsimony (MP). For the distance-based method (ME), each binary file of presence/absence is converted
into a distance matrix with the program restdist using the Nei and Li distance (1979). The distance matrix obtained is used as input for the program
fitch to infer the phylogeny under the ME algorithm. For the character based method (MP) the presence/absence matrix is used directly as input for
the program pars.
The user is allowed to generate the desired number of bootstrap pseudo-samples from the original data. To do that the program seqboot from the
PHYLIP package is used. The obtained matrices are used to reconstruct the phylogenies with the same programs as before. A consense tree for MP and
other for ME are obtained with the program consense with the minimum cut-off value indicated by the user.
The phylogenetic inference can also be done with DNA sequences, which means that sequences of 10000 nucleotides are simulated with Seq-Gen
and used directly to infer the tree. In this case the program dnadist is used to compute the distance matrix needed to reconstruct the phylogeny
under the ME method with the program fitch. To obtain the MP tree the program dnapars is used.
Each estimated tree is then compared with the reference tree using the program Ktreedist (Soria-Carrasco et al., 2007), which takes into
account both topology and branch length information of a phylogenetic tree. This program computes a K-score that measures overall differences
in the relative branch length and topology of two phylogenetic trees after scaling one of the trees to have a global divergence as similar as possible
to the other tree. The program also computes the symmetric difference or Robinson-Foulds (R-F) distance (Robinson and Foulds, 1981), which only takes
into account the topology of the phylogenetic trees.
AFLPMax generates a table for each analysis and inference method showing the mean and standard error values obtained for the K-score and
the R-F distance.
The outputs are two tables for each analysis, one corresponding to the ME results and another corresponding to the MP ones. Those tables are
displayed in the corresponding tabs of the interface but are also saved in your computer as html files in a folder under your election.
The names of the files start with ME_ or MP_ indicating the phylogenetic method used, the next part of the name refers to the analysis done
(BAND: AFLP bands, BOOT: AFLP bands with bootstrapping or SEQ: DNA sequences). The last part of the name, RES, is common to the three possible
analyses and is just to indicate that the file contains the table of results. For example, a file called ME_BAND_RES.html would contains the results
of the AFLP analysis reconstructing the phylogenies with the ME inference method.
Below you can find some examples of the tables of results, depending on the analysis done. If you run AFLPMax selecting AFLP analysis
the output will show the number of bands used to reconstruct the phylogeny on the first column and the mean and standard error obtained
from the replicates for the K-score and the R-F distance on the following ones. Values of K-score and R-F distance equal to zero indicate
perfect match between the reference tree and the AFLP-reconstructed tree.
| num band || mean K || se K || mean RF || se RF |
| 100 || 0.005794 || 0.000518 || 1.400000 || 0.221108 |
| 200 || 0.005521 || 0.000611 || 1.300000 || 0.260342 |
| 300 || 0.005034 || 0.000561 || 0.400000 || 0.221108 |
| 400 || 0.004514 || 0.000492 || 0.400000 || 0.266667 |
| 500 || 0.004157 || 0.000365 || 0.200000 || 0.200000 |
| 600 || 0.004545 || 0.000411 || 0.100000 || 0.100000 |
| 700 || 0.004402 || 0.000379 || 0.000000 || 0.000000 |
| 800 || 0.004186 || 0.000247 || 0.000000 || 0.000000 |
| 900 || 0.004398 || 0.000294 || 0.000000 || 0.000000 |
| 1000 || 0.004269 || 0.000356 || 0.000000 || 0.000000 |
| num band || mean K || se K || mean RF || se RF |
| 100 || 0.005059 || 0.000611 || 2.400000 || 0.498888 |
| 200 || 0.004461 || 0.000487 || 1.000000 || 0.447214 |
| 300 || 0.003799 || 0.000451 || 0.200000 || 0.200000 |
| 400 || 0.002553 || 0.000174 || 0.600000 || 0.305505 |
| 500 || 0.002424 || 0.000235 || 0.600000 || 0.305505 |
| 600 || 0.002713 || 0.000345 || 0.000000 || 0.000000 |
| 700 || 0.002361 || 0.000337 || 0.000000 || 0.000000 |
| 800 || 0.002071 || 0.000205 || 0.000000 || 0.000000 |
| 900 || 0.001942 || 0.000224 || 0.000000 || 0.000000 |
| 1000 || 0.002322 || 0.000320 || 0.200000 || 0.200000 |
This program AFLPMax is a Java interface to launch another programs (see above). Note that all programs launched by AFLPMax retain
their original license.