Loading and Manipulating DataThe first command that you will need to use MrBayes is$$execute$.$ This command loads information $($either a data matrix or a set of MrBayes commands$)$ from a file. You should take the file you downloaded above $($mrbayes$.$nex$)$ and move it into the MrBayes folder, then type:MrBayes$>$execute$$primates$.$nex$$NOTE:$$ there are many more commands and options available in MrBayes than we will explore in this homework if you like, feel free to check out the MrBayes Manual online!$$Specifying a Model of Sequence EvolutionBasic Model ManipulationsMost model specification is done using the$$lset$$command$.$ The most common model specifications performed in$$lset$$involve the number of substitution types and the model of rate variation across sites.$$ We'll use the Kimura $2-$parameter model $($somewhat outdated, but you are familiar with how it works!$)$:MrBayes$>$lset$$nst$=2$PartitioningIt is now widely recognized that the evolutionary process has not been homogeneous across sites. One approach to accommodating heterogeneity in the evolutionary process across sites is to divide sites into distinct subsets $($a process called partitioning$)$ and model the evolution of each subset using an independent Markov model of nucleotide substitution.The first step in performing a partitioned analysis is to define the distinct subsets of your data. These definitions are made using the charset$$command$.$ The syntax for specifying sites with$$charset$$is the same as with the$$include$$and$$exclude$$commands$.$ For instance, to specify a subset corresponding to the first codon position, type:MrBayes$>$charset$$cod$.$pos$\mathrm{.}1=1-.\backslash 3$Subsets corresponding to codon positions $2$ and $3$ could be similarly specified as:MrBayes$>$charset$$cod$.$pos$\mathrm{.}2=2-.\backslash 3$MrBayes$>$charset$$cod$.$pos$\mathrm{.}3=3-.\backslash 3$Note that each site must be assigned to one and only one subset to perform a partitioned analysis.Once all of the appropriate character sets have been defined, they need to be explicitly combined into a partition. Appropriately, this is done with a command called$$partition$.$ For instance, to specify a partitioning scheme with $3$ subsets corresponding to the three codon positions, typeMrBayes$>$partition$$cod$.$pos$=3$:cod.pos$\mathrm{.}1,$cod.pos$\mathrm{.}2,$cod.pos$\mathrm{.}3$The number immediately preceding the $"$:$"$ gives the number of subsets in the partioning scheme.After defining a partitioning scheme, you need to tell MrBayes that you actually want to use that scheme. This is done with the$$setcommand$.$ So$,$ to use the "cod.pos" partitioning scheme, typeMrBayes$>$set$$partition$=$cod.posNow that the partitioning scheme is all set, you need to define models of evolution for each subset. As above, this is done with the lset$$command again. The one additional consideration is that you need to tell MrBayes to which subsets you want any particular$$lset$$call to apply. To apply one$$lset$$command to all subsets within a partitioning scheme, typeMrBayes$>$lset$$applyto$=($all$)$nst$=2$To make sure the substitution parameters $($called revmat by MrBayes$)$ are estimated independently for each partition, you'll need to unlink the partitions.$$ You'll have to do the same for the transition$/$transversion ratio $($tratio$).$ Luckily you can do this all at once:MrBayes$>$unlink$$revmat$=($all$)$tratio$=($all$)$Note:$$ You can also provide the parameters manually rather than allowing MrBayes to calculate them we'll let MrBayes do the work this time.$$Now let's start the MCMC $($Markov Chain Monte Carlo$)$ process running.$$ Since we'll be using the default parameters $($which you can see if you type help mcmcp$),$ all you have to do is:$$MrBayes$>$mcmc$$MrBayes will start computing depending on your hardware, this may take some time $($e$.$g$.,8$ minutes on my Mac$).$ If you plan on doing this sort of thing regularly, you should look into the mpi version of MrBayes it will execute on more cores.$$At some point the SD $($standard deviation$)$ of the frequencies for each partition will drop below the threshold of $0\mathrm{.}01$; MrBayes will then ask if you would like to continue.$$ The algorithm may be able to better refine the model if you say YES, but in general this is a diminishing returns thing, and for our purposes you can simply say NO$.$ At this point the computational part of the analysis is done.$$OUTPUT ANALYSIS:$$In your MrBayes directory you should now have several new files:$$mrbayes$.$nex. run$1.$pmrbayes.nex.run$1.$tmrbayes.nex.run$2.$pmrbayes.nex.run$2.$t$$These correspond to parameter $($p$)$ and tree $($t$)$ outputs.$$Let$\text{'}$s summarize the values:$$MrBayes$>$sumpMrBayes$>$sumt$$You can read the output MrBayes provides regarding the statistics at your leisure; more interesting will be the tree "images" it generates:$$The first tree is a cladogram with the posterior probabilities for each split.$$The second tree is a phylogram with mean branch lengths represented proportionally $($so that the distance from one taxa to another can be seen$).$The trees are also printed to a file $($mrbayes$.$nex.con.tre$)$ that can be read by various tree drawing programs to generate a "prettier" output image $($say$,$ for publication$).$ If you like you can play around with that.$$ For now, complete the following questions then submit this homework to the class:$1.$ Paste the images of the trees you generated here $(15$ pts each$)2.$ Discuss the Clade Credibility Values you saw in your tree what do they lead you to believe about your tree? $(15$ pts$)$