A Practical Manual for
Real-World Phylogenetic Systematic Methods
As of 4-24-18
Handling large data sets
TnT 1.5 application: We have only been able to perform the example of phylogenetic analyses that we have manually because they involve very few possible tree arrangements. With four in-group taxa, twelve trees would have been possible. With five, the number would be sixty, and so on. For practical purposes, meaningful phylogenetic analyses can only be performed using computer algorithms. This exercise is to familiarize you with the free phylogenetic applications "TnT 1.5." (aka "the New Technologies).
You will definitely want the menu-interface PC version of the application. (An intimidating command-line only version exists for the Mac.) This can be downloaded for free from http://www.lillo.org.ar/phylogeny/tnt/ Scroll to the "download" section and click "Win (menu interface)." The licensing agreement that appears when you first use it merely requires that you not disseminate your copy and that you cite it in any publication that may result from your research using it. For best success, use a Windows 7 machine.
TnT. tutorial: On the same web page, you can also download a PowerPoint tutorial from the link in the "documentation" paragraph. This is intended for high-level users of phylogenetic software, but some of it will be useful to you.
Creating a simple data matrix using TnT.
You will have created a taxon character matrix according to the file format I provided previously using some other software (Excel, TextEdit, or WrodPad). You must follow this format precisely. However you create the file, open it in a text editor application and visually inspect it. TnT. appears to be able to read .txt files. Your file should look like this:
- If TnT. can't open your file, make sure it is simple text without tab delimitations. If you created the text file in Excel, you may need to write in the first line, "xread" manually in a text editor.
- Be sure the apostrophes you use in line two are "typewriter style" straingt up and down apostrophes, not "typeset style" curly apostrophes.
'Your file name here'
Trouble-shooting: Things I have come across:
Performing a True Phylogenetic Analysis using TnT.: Finding the most parsimonious tree
In TnT., you will be using three different forms of information:
- The data matrix you just created
- The "tree files" that TnT. will record showing your most parsimonious trees.
- The display buffer through which TnT. give a running commentary on the tasks it performs.
- Select the "Collapsing rules" option of the "Settings" menu, then click the option "None."
- Select "Open input file" from the "File" menu.
- In the pull-down menu next to the space for "file name" at the lower right of the dialog box, select "ALL files."
- Now, select the .tnt data file you created.
- A warning dialog box might ask if you really want to start working with a new data file. Click "yes."
- Under the "memory" option of the "settings" menu, increase the number of "maxtrees" from 100 to some improbably large number like 10,000.
- In the "data" menu select "outgroup taxon." Pick the taxon you want to be the outgroup.
- In the "Analyze" menu select "traditional analysis." (TnT. has powerful algorithms for running very large data sets, but we don't need them.)
- Accept all of the defaults in the "starting trees" dialog box that pops up by clicking, "search."
Output file: TnT. doesn't automatically save the information in the display buffer to a file. Ask it to do this by going to "Open output file" under the "file" menu. Name your output file with a .out extension.
Settings: TnT.'s default settings actually don't work for us. Before you proceed, you must make an adjustment:
Input file: Now you are ready to have TnT. read your data matrix. :
The display buffer should show that new data has been read. Now you need to make a second adjustment:
Designate the out-group:
Run the analysis:
Best score [TBR]: 5. 1 trees retained.
In this case, one most parsimonious tree has been identified with five state changes. The analysis took 0.06 seconds.
Viewing the Results
- In the "tree" menu select "view."
- In the "Format" menu, select "show node numbers."
- Select the "Tree Save file" option of the "File" menu.
- Select "Open, compact mode" from the options.
- Name your tree file with a .ctf extension.
- select the "read compact tree file" option of the "file" menu.
- type in (!) the name of the file you want to see, along with the .ctf extension.
- In the "optimize" menu select "synapomorphies," then "list synapomorphies."
- In the "format" menu make sure that the "map characters in color," and "preview trees" options are selected.
- Then display the trees using the "view" command of the "trees" window as before.
- With the trees displayed, go to the "characters" option of the "optimize" menu and select "character mapping." Click "select trees" in the dialog box that appears, then select the tree you want to see by moving it to the "trees included" box of the next window to appear. The tree then appears with character states color-coded. Some useful keyboard commands include:
- <enter> to advance to the next character
- <backspace> to return to the previous character
- <F4> to increase the length of the branches
- <F3> to reduce them
- <S> to save the current tree onto the display buffer
- <P> to print the current tree
- <esc> to return to the display buffer
You will need a convenient way of referring to nodes in the trees:
Save trees: You will want to save your set of most parsimonious trees so that you can reopen them later, especially if your analysis took lots of time. To do that:
Opening trees: To reopen your tree file later:
View character states on trees:
Consensus trees: If you have a great many most parsimonious trees, viewing them all can be very tedious. To make things easy, we usually summarize tree information on consensus trees, which show the information on which the set of most parsimonious trees agrees. There are several consensus algorithms, but we will use the two we learned about in class:
- Select the "consensus" option of the "tree" menu. In the consensus dialog box, select "strict" and "save to RAM" options, then click "OK."
- Again, select the "consensus" option of the "tree" menu. In the consensus dialog box, select "Majority Rule" and "save to RAM" options, then click "OK."
Majority rule consensus:
Use the same character mapping method described above to see character state changes on your consensus trees.
Things I have come across:
- The currently available version of TnT. is definitely compatible with machines running Windows 7. If the application fails to launch, it is probably because of an operating system incompatibility. If you are using a Windows NT machine, you may have trouble.
- If your analysis returns only a single most parsimonious tree when you suspect there should be many, double check that you have clicked the "none" option from the "collapse rules" item of the "settings" menu.
- For reasons I can't fathom, the <S> command to save a tree to the display buffer doesn't work reliably. You may need to get screen shots of the trees for inclusion in your final write-up.
This concludes our introductory demo. There is, of course, much much more to phylogenetic analysis, but now you should be able to explore it on your own. If you have digested the information here, however, you can perform a competent phylogenetic analysis.