Difference between revisions of "Phylogenetics: Large Scale Maximum Likelihood Analyses"
Paul Lewis (Talk | contribs) (→Running RAxML on the cluster) |
(→Comparing GARLI and RAxML) |
||
(29 intermediate revisions by 2 users not shown) | |||
Line 2: | Line 2: | ||
|- | |- | ||
|rowspan="2" valign="top"|[[Image:Adiantum.png|200px]] | |rowspan="2" valign="top"|[[Image:Adiantum.png|200px]] | ||
− | |<span style="font-size: x-large">[http:// | + | |<span style="font-size: x-large">[http://phylogeny.uconn.edu/courses/ EEB 5349: Phylogenetics]</span> |
|- | |- | ||
| This lab explores two programs (GARLI and RAxML) designed specifically for maximum likelihood analyses on a large scale (hundreds of taxa). | | This lab explores two programs (GARLI and RAxML) designed specifically for maximum likelihood analyses on a large scale (hundreds of taxa). | ||
− | |||
− | |||
|} | |} | ||
Line 39: | Line 37: | ||
ofprefix = 50 | ofprefix = 50 | ||
The ofprefix is used by GARLI to begin the name of all output files. I usually use something ''different'' than the data file name here. That way, if you eventually want to delete all of the various files that GARLI creates, you can just say <tt>rm -f 50*</tt> ("remove all files beginning "50", the -f means "force" i.e. don't ask if it is ok, just do it) without wiping out your data file as well! (Sounds like the voice of experience, doesn't it?!) | The ofprefix is used by GARLI to begin the name of all output files. I usually use something ''different'' than the data file name here. That way, if you eventually want to delete all of the various files that GARLI creates, you can just say <tt>rm -f 50*</tt> ("remove all files beginning "50", the -f means "force" i.e. don't ask if it is ok, just do it) without wiping out your data file as well! (Sounds like the voice of experience, doesn't it?!) | ||
+ | |||
+ | '''Do only one search replicate''' | ||
+ | searchreps = 1 | ||
'''Specify no invariable sites''' | '''Specify no invariable sites''' | ||
invariantsites = none | invariantsites = none | ||
This will cause GARLI to use the GTR+G model rather than the GTR+I+G model, which will facilitate comparisons with RAxML. | This will cause GARLI to use the GTR+G model rather than the GTR+I+G model, which will facilitate comparisons with RAxML. | ||
− | |||
− | |||
− | |||
Save the <tt>garli.conf</tt> file when you have made these changes. | Save the <tt>garli.conf</tt> file when you have made these changes. | ||
Line 51: | Line 49: | ||
=== The tip of the GARLI iceberg === | === The tip of the GARLI iceberg === | ||
− | As you can see from the number of entries in the control file, we are not going to learn all there is to know about GARLI in one lab session. One major omission is any discussion about bootstrapping, which is very easy to do in GARLI: just set <tt>bootstrapreps</tt> to some number other than 0 (e.g. 100) in your <tt>garli.conf</tt> file. I encourage you to download and read the excellent GARLI manual, especially if you want to use amino acid or codon models. | + | As you can see from the number of entries in the control file, we are not going to learn all there is to know about GARLI in one lab session. One major omission is any discussion about bootstrapping, which is very easy to do in GARLI: just set <tt>bootstrapreps</tt> to some number other than 0 (e.g. 100) in your <tt>garli.conf</tt> file ('''but don't do this today'''). I encourage you to download and read the excellent GARLI manual, especially if you want to use amino acid or codon models. |
=== Log into the cluster === | === Log into the cluster === | ||
Line 60: | Line 58: | ||
=== Create a folder and a script for the run === | === Create a folder and a script for the run === | ||
− | Create a directory named <tt> | + | Create a directory named <tt>garliraxml</tt> inside your home directory and use your favorite file transfer method (scp, psftp, Fugu, FileZilla, etc.) to get <tt>garli.conf</tt> into that directory. |
− | Now download the data file into the <tt> | + | Now download the data file into the <tt>garliraxml</tt> directory: |
− | curl http://hydrodictyon.eeb.uconn.edu/ | + | curl http://hydrodictyon.eeb.uconn.edu/people/plewis/courses/phylogenetics/data/rbcL50.nex > rbcL50.nex |
− | Finally, create the script file you will hand to the <tt>qsub</tt> command to start the run. Use the nano editor to create a file named <tt>gogarli</tt> inside the <tt> | + | Finally, create the script file you will hand to the <tt>qsub</tt> command to start the run. Use the nano editor to create a file named <tt>gogarli</tt> inside the <tt>garliraxml</tt> directory with the following contents: |
#$ -cwd | #$ -cwd | ||
#$ -S /bin/bash | #$ -S /bin/bash | ||
Line 75: | Line 73: | ||
garli garli.conf | garli garli.conf | ||
− | The lines starting with <tt>#$</tt> represent commands that qsub understands. Any line starting with a hash (#) is interpreted as a comment by all linux/unix interpreters, but the extra dollar sign ($) immediately after the hash is used by the qsub program as a flag, telling it that what follows should be interpreted as a command. All of these could be entered on the qsub command line as well, but it is to forget something unless you put it in a script like this. | + | The lines starting with <tt>#$</tt> represent commands that qsub understands. Any line starting with a hash (#) is interpreted as a comment by all linux/unix interpreters, but the extra dollar sign ($) immediately after the hash is used by the qsub program as a flag, telling it that what follows should be interpreted as a command. All of these could be entered on the qsub command line as well, but it is easy to forget something unless you put it in a script like this. |
− | The command <tt>#$ -cwd</tt> tells qsub to put output files in the current working directory, the directory in which it was invoked (which will be the <tt> | + | The command <tt>#$ -cwd</tt> tells qsub to put output files in the current working directory, the directory in which it was invoked (which will be the <tt>garliraxml</tt> directory). |
The command <tt>#$ -S /bin/bash</tt> tells qsub to use the bash program to interpret the script. You have already been using the bash program; it interprets the commands you type when you are logged into the cluster. This qsub command just tells qsub to use the same program to interpret commands in the <tt>gogarli</tt> file. | The command <tt>#$ -S /bin/bash</tt> tells qsub to use the bash program to interpret the script. You have already been using the bash program; it interprets the commands you type when you are logged into the cluster. This qsub command just tells qsub to use the same program to interpret commands in the <tt>gogarli</tt> file. | ||
Line 95: | Line 93: | ||
Here is the command to start the job: | Here is the command to start the job: | ||
qsub gogarli | qsub gogarli | ||
− | You should issue this command from inside the <tt> | + | You should issue this command from inside the <tt>garliraxml</tt> directory, which should contain 3 files: <tt>gogarli</tt>, <tt>rbcL50.nex</tt> and <tt>garli.conf</tt>. |
Check progress every few minutes using the [http://137.99.46.187/wiki/index.php/Qstat qstat] command. This run will take about 4 minutes. If you get bored, you can use this command to see the tail end of the log file that GARLI creates automatically: | Check progress every few minutes using the [http://137.99.46.187/wiki/index.php/Qstat qstat] command. This run will take about 4 minutes. If you get bored, you can use this command to see the tail end of the log file that GARLI creates automatically: | ||
Line 105: | Line 103: | ||
=== Mailing the tree to yourself === | === Mailing the tree to yourself === | ||
− | After GARLI has finished, you should download the tree file (50.best.tre) using the PSFTP get command, but here is another handy trick: you can email the tree to yourself using this command (issue this from within the | + | After GARLI has finished, you should download the tree file (50.best.tre) using the PSFTP get command, but here is another handy trick: you can email the tree to yourself using this command (issue this from within the garliraxml directory where the tree file is located): |
mail paul.lewis@uconn.edu < 50.best.tre | mail paul.lewis@uconn.edu < 50.best.tre | ||
This command will send mail to paul.lewis@uconn.edu, and the body of the email message will come from the file 50.best.tre! | This command will send mail to paul.lewis@uconn.edu, and the body of the email message will come from the file 50.best.tre! | ||
Line 112: | Line 110: | ||
=== Files produced by GARLI === | === Files produced by GARLI === | ||
− | After your run finishes, you should find these files in your <tt> | + | After your run finishes, you should find these files in your <tt>garliraxml</tt> folder. Download them to your laptop and view them to answer the questions: |
==== <tt>50.screen.log</tt> ==== | ==== <tt>50.screen.log</tt> ==== | ||
Line 166: | Line 164: | ||
=== Running RAxML on the cluster === | === Running RAxML on the cluster === | ||
− | Hopefully, you have created the <tt>rbcL50.dat</tt> file in your <tt> | + | Hopefully, you have created the <tt>rbcL50.dat</tt> file in your <tt>garliraxml</tt> directory. If not, go ahead and move it there. Then use nano to create a <tt>gorax</tt> script file that contains the following: |
#$ -cwd | #$ -cwd | ||
#$ -S /bin/bash | #$ -S /bin/bash | ||
Line 181: | Line 179: | ||
The last line invokes raxml and requires the most explanation. First, RAxML does not use a control file like GARLI, so all options must be specified on the command line when it is invoked. Let's take each option in turn: | The last line invokes raxml and requires the most explanation. First, RAxML does not use a control file like GARLI, so all options must be specified on the command line when it is invoked. Let's take each option in turn: | ||
− | * '''-p 13579''' provides a pseudorandom number seed to RAxML to use when it generates its starting tree. It is a good idea to specify some number here so that you have the option of exactly recreating the analysis later. | + | * '''-p 13579''' provides a pseudorandom number seed to RAxML to use when it generates its starting tree. It is a good idea to specify some number here so that you have the option of exactly recreating the analysis later. By the way, the "p" stands for parsimony, as this seed is used to generate the random addition sequence for constructing a starting tree using parsimony. |
* '''-N 1''' tells RAxML to just perform one search replicate. | * '''-N 1''' tells RAxML to just perform one search replicate. | ||
* '''-e 0.00001''' sets the precision with which model parameters will be estimated. RAxML will search for better combinations of parameter values until it fails to increase the log-likelihood by more than this amount. Ordinarily, the default value (0.1) is sufficient, but we are making RAxML work harder so that the results are more comparable to GARLI, which does a fairly thorough final parameter optimization. | * '''-e 0.00001''' sets the precision with which model parameters will be estimated. RAxML will search for better combinations of parameter values until it fails to increase the log-likelihood by more than this amount. Ordinarily, the default value (0.1) is sufficient, but we are making RAxML work harder so that the results are more comparable to GARLI, which does a fairly thorough final parameter optimization. | ||
− | * '''-m GTRCAT''' tells RAxML to use the GTR+CAT model for the search | + | * '''-m GTRCAT''' tells RAxML to use the GTR+CAT model for the search. I will explain in lecture how the CAT model works in RAxML - it is essentially a sites-specific-rates model for rate heterogeneity, but derives the categories and relative rates from the data set itself before the search begins. |
* '''-s rbcL50.dat''' provides the name of the data file. | * '''-s rbcL50.dat''' provides the name of the data file. | ||
* '''-n BASIC''' supplies a suffix to be appended to all output file names | * '''-n BASIC''' supplies a suffix to be appended to all output file names | ||
− | Start the run by entering this from within the <tt> | + | Start the run by entering this from within the <tt>garliraxml</tt> directory (where your <tt>rbcL50.dat</tt> and <tt>gorax</tt> files are located): |
qsub gorax | qsub gorax | ||
=== Bootstrapping with RAxML === | === Bootstrapping with RAxML === | ||
− | After your first RAxML run finishes (probably within | + | After your first RAxML run finishes (probably within a minute), start a second, longer run to perform bootstrapping. Modify the last line of your <tt>gorax</tt> file as follows: |
− | # raxml -p 13579 -N 1 -e 0.00001 -m | + | # raxml -p 13579 -N 1 -e 0.00001 -m GTRCAT -s rbcL50.dat -n BASIC |
raxml -f a -x 12345 -p 13579 -N 100 -m GTRCAT -s rbcL50.dat -n FULL | raxml -f a -x 12345 -p 13579 -N 100 -m GTRCAT -s rbcL50.dat -n FULL | ||
Line 204: | Line 202: | ||
This file contains some basic information about the run. Use this file to answer these questions: | This file contains some basic information about the run. Use this file to answer these questions: | ||
− | <div style="background-color: #ddddff">How long did RAxML require to perform 100 bootstrap replicates? | + | <div style="background-color: #ddddff">How long did RAxML require to perform 100 bootstrap replicates? {{title|116.512911 seconds|Paul's results}}</div> |
− | + | <div style="background-color: #ddddff">What is the log-likelihood of the best tree found by RAxML? {{title|-18329.025632|Paul's results}}</div> | |
− | + | ||
− | + | ||
− | <div style="background-color: #ddddff">What is the log-likelihood of the best tree found by RAxML?</div> | + | |
==== <tt>RAxML_bestTree.FULL</tt> ==== | ==== <tt>RAxML_bestTree.FULL</tt> ==== | ||
Line 222: | Line 217: | ||
This file contains the best tree with bootstrap support values embedded in the tree description. Load this tree into FigTree. FigTree will ask you what name you want to use for the support values. Pick a name such as "bootstraps" and click Ok. Once the tree is visible, check ''Node labels'' on the left, chooose "bootstraps" (or whatever you named them) from the ''Display'' list, and increase the font size so you can see it (ok, you are probably young enough that you can still see the numbers without magnification!). | This file contains the best tree with bootstrap support values embedded in the tree description. Load this tree into FigTree. FigTree will ask you what name you want to use for the support values. Pick a name such as "bootstraps" and click Ok. Once the tree is visible, check ''Node labels'' on the left, chooose "bootstraps" (or whatever you named them) from the ''Display'' list, and increase the font size so you can see it (ok, you are probably young enough that you can still see the numbers without magnification!). | ||
− | <div style="background-color: #ddddff">Can you tell by viewing this tree in FigTree that this is ''not'' the bootstrap majority-rule consensus tree?</div> | + | <div style="background-color: #ddddff">Can you tell by viewing this tree in FigTree that this is ''not'' the bootstrap majority-rule consensus tree? What evidence did you use? {{title|some support values are less than 50, which would not be the case for a majority rule tree|answer}}</div> |
− | + | We could have instructed RAxML to create the majority rule tree using the <tt>"-J MR</tt> option, but by default it maps bootstrap support values onto the splits of the maximum likelihood tree topology. | |
− | + | <!-- | |
+ | == Computing the RAxML majority rule consensus tree using PAUP* == | ||
+ | |||
+ | RAxML performed bootstrapping but did not provide a majority rule consensus tree. Although RAxML could have been asked to create the majority rule tree (using the <tt>"-J MR</tt> option), I did not have you do this so that I could show you how to do it in PAUP (this is a useful thing to know how to do). RAxML created the file <tt>RAxML_bootstrap.FULL</tt> containing the best tree from each of the 100 bootstrap replicates. Your goal in this section is to get PAUP to create the majority rule consensus tree from this file. To do this, create a nexus file named <tt>majrule.nex</tt> containing this text: | ||
#nexus | #nexus | ||
begin paup; | begin paup; | ||
− | + | exe rbcL50.nex; | |
+ | gettrees file=RAxML_bootstrap.FULL; | ||
+ | contree all / nostrict majrule treefile=raxmlmajrule.tre; | ||
end; | end; | ||
+ | PAUP requires that you read in the data file before reading in the trees (so that it has a list of valid taxon names), hence the initial <tt>exe</tt> command. The <tt>gettrees</tt> command should be straightforward: this just tells PAUP to read in all the trees in the file <tt>RAxML_bootstrap.FULL</tt>. The last <tt>contree</tt> command instructs PAUP to create a majority rule consensus tree (<tt>majrule</tt>) from the trees it now has in memory, but not to create the strict consensus tree (<tt>nostrict</tt>) that it would ordinarily do by default. Finally, the <tt>treefile</tt> command provides the name of a file in which to save the majority rule consensus tree. As always, if you want to see all the options for a command like <tt>contree</tt>, just put a question mark after the command name in PAUP (and use the <tt>help</tt> command to get a list of all commands): | ||
+ | paup> contree ?; | ||
− | + | Usage: ConTree [tree-list] [/ options...] ; | |
− | + | ||
− | + | ||
− | + | ||
− | begin paup; | + | Available options: |
− | + | ||
− | + | Keyword ------- Option type --------------------- Current setting ---------- | |
− | + | strict no|yes no | |
− | + | semistrict no|yes no | |
− | + | adams no|yes no | |
+ | majRule no|yes yes | ||
+ | percent <integer-value> 50 | ||
+ | LE50 no|yes no | ||
+ | useTreeWts no|yes no | ||
+ | grpFreq no|yes yes | ||
+ | indices no|yes no | ||
+ | showTree no|yes yes | ||
+ | treeFile <tree-file-name> <none> | ||
+ | saveSupport no|brlens|nodeLabels|Both nodeLabels | ||
+ | replace no|yes *no | ||
+ | append no|yes *no | ||
+ | tCompress no|yes no | ||
+ | rootMethod outgroup|lundberg|midpoint outgroup | ||
+ | outRoot polytomy|paraphyl|monophyl polytomy | ||
+ | forcePolyt no|yes no | ||
+ | *Option is nonpersistent | ||
+ | Run the <tt>majrule.nex</tt> in PAUP as follows (no need to qlogin for this, but do exit PAUP if you it is still running): | ||
+ | paup majrule.nex | ||
+ | and note that the majority rule tree was saved in the file <tt>raxmlmajrule.tre</tt>. | ||
+ | --> | ||
+ | |||
+ | == Comparing GARLI and RAxML == | ||
+ | |||
+ | To compare the two programs, use the nano editor to create a tree file named <tt>combined.nex</tt> and specify the best tree from both programs and a paup block to compute the likelihoods of these two trees under the GTR+G model. Your <tt>combined.nex</tt> command block should contain the following: | ||
+ | |||
+ | #nexus | ||
+ | begin paup; | ||
+ | exe rbcL50.nex; | ||
+ | gettrees file=RAxML_bestTree.FULL; | ||
+ | deroot; | ||
+ | gettrees file=50.best.tre mode = 7; | ||
+ | set criterion=likelihood; | ||
+ | lset nst=6 rmatrix=estimate rates=gamma shape=estimate; | ||
+ | lscores all; | ||
+ | agree all; | ||
+ | end; | ||
+ | |||
+ | Here, <tt>gettrees</tt> command reads the two tree files (RAxML_bestTree.FULL, and 50.best2.tre) you generated using the programs RaxML and Garli. The <tt>mode=7</tt> tells PAUP to keep the RaxML tree (RAxML_bestTree.FULL) in the memory while loading the Garli tree (50.best.tre). You should have two tree files and a data file (rbcL50.nex) in the same folder to execute combined.nex. The comment [&U] in the Garli tree file tells PAUP that the tree is unrooted. Therefore, to compare it with RAxML tree (which does not contain any rooting information in the file) you can add the command <tt>deroot</tt> (after gettrees file=RAxML_bestTree.FULL;). | ||
+ | |||
+ | Which tree has the best log-likelihood? You will probably find that GARLI's likelihood is slightly different than RAxML, but both will be nearly the same value. Both of these approaches will give different answers if you run them multiple times under different random number seeds, so you should probably do several replicates (or a bootstrap analysis) before making anything too momentous of the results. | ||
− | + | The last command given to PAUP was "agree all". This computes an agreement subtree from the results of the two analyses. | |
− | + | <div style="background-color: #ddddff">Can you figure out from PAUP's output how many taxa (out of the 50 total) it had to omit in order to find an agreement subtree? {{title|5 taxa were deleted, 45 taxa are in the agreement subtree|answer}}</div> | |
[[Category:Phylogenetics]] | [[Category:Phylogenetics]] |
Revision as of 22:25, 17 February 2016
EEB 5349: Phylogenetics | |
This lab explores two programs (GARLI and RAxML) designed specifically for maximum likelihood analyses on a large scale (hundreds of taxa). |
Contents
- 1 The data
- 2 Part A: Starting a GARLI run on the cluster
- 3 Part B: Starting a RAxML run on the cluster
- 4 Comparing GARLI and RAxML
The data
The data that we will use today comprises 50 rbcL sequences from various green plants: 5 green algae, 5 bryophytes (mosses, liverworts, hornworts), 5 ferns, 9 gymnosperms, and 26 angiosperms (flowering plants; including some you will recognize such as oregano, coffee, tea, spinach, poison ivy, rice and chocolate). The tree on the right shows the tree I obtained from GARLI using the GY94 codon model, color-coded so that bryophytes are yellow, ferns are green, gymnosperms are blue and angiosperms are pink. This history of green plants shows several key innovations: embryos are what gave the first bryophytes an edge over aquatic algae on land; branched sporophytes and vascular tissues allowed the first ferns to grow taller and disperse more spores compared to their bryophyte ancestors; seeds and pollen were the big inventions that led to the rise of gymnosperms; and of course flowers allowed efficient pollination by insects and led the the diversification of the angiosperms.Part A: Starting a GARLI run on the cluster
GARLI (Genetic Algorithm for Rapid Likelihood Inference) is a program written by Derrick Zwickl for estimating the phylogeny using maximum likelihood, and is currently one of the best programs to use if you have a large problem (i.e. many taxa). GARLI now (as of version 2.01) gives you considerable choice in substitution models: GTR[+I][+G] or codon models for nucleotides, plus several choices for amino acids. The genetic algorithm (or GA, for short) search strategy used by GARLI is like other heuristic search strategies in that it cannot guarantee that the optimal tree will be found. Thus, as with all heuristic searches, it is a good idea to run GARLI several times (using different pseudorandom number seeds) to see if there is any variation in the estimated tree. By default, GARLI will conduct two independent searches. If you have a multicore processor (newer Intel-based Macs and PCs are duo-core), GARLI can take advantage of this and use all of your CPUs simultaneously.
Today you will run GARLI on the cluster for a dataset with 50 taxa. This is not a particularly large problem, but has the advantage that you will be able to analyze it several times using both GARLI and RAxML within a lab period. Instead of each of us running GARLI several times, we will each run it once and compare notes at the end of the lab.
Preparing the GARLI control file
Like many programs, GARLI uses a control file to specify the settings it will use during a run. Most of the default settings are fine, but you will need to change a few of them before running GARLI.
Obtain a copy of the control file
The first step is to obtain a copy of the GARLI default control file. Go to the GARLI download page and download a version of GARLI appropriate for your platform (Mac or Windows). The only reason you are downloading GARLI is to obtain a copy of the default control file; you should use the cluster for all your GARLI runs.
Once you have downloaded and unpacked GARLI on your computer, make a copy of the garli.conf.nuc.defaultSettings file and rename the copy garli.conf, then open it in your text editor.
Editing garli.conf
You will only need to change four lines.
Specify the data file name (note the capital L)
datafname = rbcL50.nex
Specify the prefix for output files
ofprefix = 50
The ofprefix is used by GARLI to begin the name of all output files. I usually use something different than the data file name here. That way, if you eventually want to delete all of the various files that GARLI creates, you can just say rm -f 50* ("remove all files beginning "50", the -f means "force" i.e. don't ask if it is ok, just do it) without wiping out your data file as well! (Sounds like the voice of experience, doesn't it?!)
Do only one search replicate
searchreps = 1
Specify no invariable sites
invariantsites = none
This will cause GARLI to use the GTR+G model rather than the GTR+I+G model, which will facilitate comparisons with RAxML.
Save the garli.conf file when you have made these changes.
The tip of the GARLI iceberg
As you can see from the number of entries in the control file, we are not going to learn all there is to know about GARLI in one lab session. One major omission is any discussion about bootstrapping, which is very easy to do in GARLI: just set bootstrapreps to some number other than 0 (e.g. 100) in your garli.conf file (but don't do this today). I encourage you to download and read the excellent GARLI manual, especially if you want to use amino acid or codon models.
Log into the cluster
Log into the cluster using the command:
ssh bbcsrv3.biotech.uconn.edu
Go back to the Phylogenetics: Bioinformatics Cluster lab if you've forgotten some details.
Create a folder and a script for the run
Create a directory named garliraxml inside your home directory and use your favorite file transfer method (scp, psftp, Fugu, FileZilla, etc.) to get garli.conf into that directory.
Now download the data file into the garliraxml directory:
curl http://hydrodictyon.eeb.uconn.edu/people/plewis/courses/phylogenetics/data/rbcL50.nex > rbcL50.nex
Finally, create the script file you will hand to the qsub command to start the run. Use the nano editor to create a file named gogarli inside the garliraxml directory with the following contents:
#$ -cwd #$ -S /bin/bash #$ -o out.txt #$ -e err.txt #$ -m ea #$ -M your.name@uconn.edu #$ -N jobname garli garli.conf
The lines starting with #$ represent commands that qsub understands. Any line starting with a hash (#) is interpreted as a comment by all linux/unix interpreters, but the extra dollar sign ($) immediately after the hash is used by the qsub program as a flag, telling it that what follows should be interpreted as a command. All of these could be entered on the qsub command line as well, but it is easy to forget something unless you put it in a script like this.
The command #$ -cwd tells qsub to put output files in the current working directory, the directory in which it was invoked (which will be the garliraxml directory).
The command #$ -S /bin/bash tells qsub to use the bash program to interpret the script. You have already been using the bash program; it interprets the commands you type when you are logged into the cluster. This qsub command just tells qsub to use the same program to interpret commands in the gogarli file.
The command #$ -o out.txt tells qsub to save any output into the file out.txt.
The command #$ -e err.txt tells qsub to save any error messages into the file err.txt.
The command #$ -m ea tells qsub to send you an email when your job ends (e) or aborts (a). The command #$ -M your.name@uconn.edu tells qsub what your email address is (be sure to specify your own email address here). This is an extremely helpful feature! This feature may not work if you provide a non-uconn email address (feel free to try your gmail address and let me know if it works).
Finally, the -N command provides a job name that qsub will use to identify your job. This is entirely optional, but I find it very helpful to give short names to my runs so that when I get the email I know which run has finished (most useful if you have several runs going simultaneously). Be sure to replace "jobname" with something more meaningful (keep your job name simple - no embedded spaces or punctuation, and you will find that it is better to use 6 characters or fewer, as longer job names get truncated in qstat output).
The last line starts garli.
Submit the job
Here is the command to start the job:
qsub gogarli
You should issue this command from inside the garliraxml directory, which should contain 3 files: gogarli, rbcL50.nex and garli.conf.
Check progress every few minutes using the qstat command. This run will take about 4 minutes. If you get bored, you can use this command to see the tail end of the log file that GARLI creates automatically:
tail 50.log00.log
The tail command is like the cat command except that it only shows you the last few lines of the file (which often is just what you need). If you need to see more lines at the end of the file, you can specify the number using the -n option:
tail -n 100 50.log00.log
The above command will show the last 100 lines.
Files produced by GARLI
After your run finishes, you should find these files in your garliraxml folder. Download them to your laptop and view them to answer the questions:
50.screen.log
This file saves the output that would have been displayed had you been running GARLI on your laptop. (Hover over the phrase "Paul's result" below to see what values I saw in my 50.screen.log file; your run may differ because we all started with a different seed.)
50.log00.log
This file shows the best log-likelihood at periodic intervals throughout the run. It would be useful if you wanted to plot the progress of the run either as a function of time or generation.
50.best.tre
This is a NEXUS tree file that can be opened in FigTree, TreeView, PAUP*, or a number of other phylogenetic programs. Try using FigTree to open it. The best place to root it is on the branch leading to Nephroselmis. In FigTree, click this branch and use the Reroot tool to change the rooting. I also find that trees look better if you click the Order nodes checkbox, which is inside the Trees tab on the left side panel of FigTree.
Part B: Starting a RAxML run on the cluster
Another excellent ML program for large problems is RAxML, written by Alexandros Stamatakis. This program is exceptionally fast, and has been used to estimate maximum likelihood trees for 25,000 taxa! Let's run RAxML on the same data as GARLI and compare results.
Preparing the data file
While GARLI reads NEXUS files, RAxML uses a simpler format. It is easy to use the nano editor to make the necessary changes, however. First, make a copy of your rbcL50.nex file:
cp rbcL50.nex rbcL50.dat
Open rbcL50.dat in nano and use Ctrl-k repeatedly to remove these initial lines:
#nexus begin data; dimensions ntax=50 nchar=1314; format datatype=dna gap=- missing=?; matrix
Add a new first line to the file that looks like this:
50 1314
Now use the down arrow to go to the end of the file and remove the last two lines (again using Ctrl-k):
; end;
Save the file using Ctrl-x and you are ready to run RAxML!
The tip of the RAxML iceberg
As with GARLI, RAxML is full of features that we will not have time to explore today. Try typing raxml -h at the linux command prompt to see a brief description of all the available options. The manual does a nice job of explaining all the features so I recommend reading it if you use RAxML for your own data.
Running RAxML on the cluster
Hopefully, you have created the rbcL50.dat file in your garliraxml directory. If not, go ahead and move it there. Then use nano to create a gorax script file that contains the following:
#$ -cwd #$ -S /bin/bash #$ -o raxout.txt #$ -e raxerr.txt #$ -m ea #$ -M your.name@uconn.edu #$ -N jobname raxml -p 13579 -N 1 -e 0.00001 -m GTRCAT -s rbcL50.dat -n BASIC
You'll note that this is similar to the gogarli script we created earlier, but please read the explanation below before submitting the run to the cluster.
The qsub command lines are the same except that we specified raxout.txt rather than out.txt and raxerr.txt rather than err.txt (this is so that our RAxML run will not overwrite the files generated by our GARLI run).
The last line invokes raxml and requires the most explanation. First, RAxML does not use a control file like GARLI, so all options must be specified on the command line when it is invoked. Let's take each option in turn:
- -p 13579 provides a pseudorandom number seed to RAxML to use when it generates its starting tree. It is a good idea to specify some number here so that you have the option of exactly recreating the analysis later. By the way, the "p" stands for parsimony, as this seed is used to generate the random addition sequence for constructing a starting tree using parsimony.
- -N 1 tells RAxML to just perform one search replicate.
- -e 0.00001 sets the precision with which model parameters will be estimated. RAxML will search for better combinations of parameter values until it fails to increase the log-likelihood by more than this amount. Ordinarily, the default value (0.1) is sufficient, but we are making RAxML work harder so that the results are more comparable to GARLI, which does a fairly thorough final parameter optimization.
- -m GTRCAT tells RAxML to use the GTR+CAT model for the search. I will explain in lecture how the CAT model works in RAxML - it is essentially a sites-specific-rates model for rate heterogeneity, but derives the categories and relative rates from the data set itself before the search begins.
- -s rbcL50.dat provides the name of the data file.
- -n BASIC supplies a suffix to be appended to all output file names
Start the run by entering this from within the garliraxml directory (where your rbcL50.dat and gorax files are located):
qsub gorax
Bootstrapping with RAxML
After your first RAxML run finishes (probably within a minute), start a second, longer run to perform bootstrapping. Modify the last line of your gorax file as follows:
# raxml -p 13579 -N 1 -e 0.00001 -m GTRCAT -s rbcL50.dat -n BASIC raxml -f a -x 12345 -p 13579 -N 100 -m GTRCAT -s rbcL50.dat -n FULL
Note that I've used a # character to comment out our previous raxml line. (Feel free to simply delete that line if you wish.) Go ahead and start this run using qsub. This one will take longer, but not as long as you might expect (about 10 minutes). It will conduct a bootstrap analysis involving 100 bootstrap replicates (-N 100) using the GTR+CAT model. The -x 12345 specifies a starting seed for the bootstrap resampling. For every 5 bootstrap replicates performed, RAxML will climb uphill on the original dataset starting from the tree estimated for that bootstrap replicate. This provides a series of searches for the maximum likelihood tree starting from different, but reasonable, starting trees. The -f a on the command line sets up this combination of bootstrapping and ML searching.
Files produced by RAxML
RAxML_info.FULL
This file contains some basic information about the run. Use this file to answer these questions:
RAxML_bestTree.FULL
This file holds the best tree found. It is not a NEXUS tree file, but simply a tree description; however, FigTree is able to open such files.
RAxML_bootstrap.FULL
This file holds the trees resulting from bootstrapping (also not NEXUS format; one tree description per line). These trees do not have branch lengths. You can open this file in FigTree and use the arrow buttons to move from one to the next.
RAxML_bipartitions.FULL
This file contains the best tree with bootstrap support values embedded in the tree description. Load this tree into FigTree. FigTree will ask you what name you want to use for the support values. Pick a name such as "bootstraps" and click Ok. Once the tree is visible, check Node labels on the left, chooose "bootstraps" (or whatever you named them) from the Display list, and increase the font size so you can see it (ok, you are probably young enough that you can still see the numbers without magnification!).
We could have instructed RAxML to create the majority rule tree using the "-J MR option, but by default it maps bootstrap support values onto the splits of the maximum likelihood tree topology.
Comparing GARLI and RAxML
To compare the two programs, use the nano editor to create a tree file named combined.nex and specify the best tree from both programs and a paup block to compute the likelihoods of these two trees under the GTR+G model. Your combined.nex command block should contain the following:
#nexus begin paup; exe rbcL50.nex; gettrees file=RAxML_bestTree.FULL; deroot; gettrees file=50.best.tre mode = 7; set criterion=likelihood; lset nst=6 rmatrix=estimate rates=gamma shape=estimate; lscores all; agree all; end;
Here, gettrees command reads the two tree files (RAxML_bestTree.FULL, and 50.best2.tre) you generated using the programs RaxML and Garli. The mode=7 tells PAUP to keep the RaxML tree (RAxML_bestTree.FULL) in the memory while loading the Garli tree (50.best.tre). You should have two tree files and a data file (rbcL50.nex) in the same folder to execute combined.nex. The comment [&U] in the Garli tree file tells PAUP that the tree is unrooted. Therefore, to compare it with RAxML tree (which does not contain any rooting information in the file) you can add the command deroot (after gettrees file=RAxML_bestTree.FULL;).
Which tree has the best log-likelihood? You will probably find that GARLI's likelihood is slightly different than RAxML, but both will be nearly the same value. Both of these approaches will give different answers if you run them multiple times under different random number seeds, so you should probably do several replicates (or a bootstrap analysis) before making anything too momentous of the results.
The last command given to PAUP was "agree all". This computes an agreement subtree from the results of the two analyses.