CSB472

Phylogenetics Tutorial

In this phylogenetics tutorial, you will:

.    Gain experience with multiple phylogenetic methods.

.    Investigate how the different MSA approaches and parameters influence the inferred evolutionary histories.

.    Investigate how using an evolutionary substitution model influences the inferred history vs. no evolutionary model.

.    Investigate how phylogenetic trees produced from aligned DNA sequences differ from those produced from aligned protein sequences.

We will be using MEGA11 for this lab, but there are many other phylogenetics tools available. As a reminder, you can access MEGA from https://www.megasoftware.net/. Note that there are some notable changes to the tree viewer in MEGA11 relative to earlier versions. I have also listed a number of alternative tools in Appendix 3.

Important - the quiz is different for this lab. You will be asked to upload a short summary of your analysis results. This is explained at the end of the lab and on the class website.

CASE 1: Exploring the Impact of Alignment Parameters

Biological Plausibility vs. Phylogenetic Support and the Impact of Alignment Methods & Parameters

Let's first look at the impact of using a codon-based alignment approach when aligning coding sequencing. Recall, when you align a coding sequence you can either align the DNA sequence directly, or you can translate the DNA to the corresponding protein sequence, perform. the amino acid-based alignment, and then back-translate the alignment to the DNA sequence. Doing this imposes a degree of biological plausibility since insertions and deletions that change the coding sequence will almost always result in a premature stop codon (i.e., a nonsense mutation). Since these mutations create truncated proteins, they should be strongly selected against, and therefore, rapidly culled from the population by natural selection. Given this, we prefer to perform. codon-based alignments, when possible, but you need to recognize that doing this imposes constraints on where gaps can be inserted. Also, I need to emphasize that this is only appropriate when working with a coding sequence. Let's see how these constraints impact our phylogenetic analysis. We will also examine how different alignment approaches and parameters influence phylogenetic results.

We will use a dataset of HopM1 type III effector proteins from the plant pathogenic bacterium Pseudomonas syringae. As discussed previously, type III effectors are translocated directly from the bacterial cell to the host cell via a type III secretion system. They have evolved in order to promote bacterial fitness (virulence in this context) but can also trigger the host immune response if the host has the appropriate immune receptors. I have provided two alignments. You will primarily work with

HopM1_clustal_10_0.1_codon.fas, while most of the analysis for HopM1_clustal  1  1.fas will be provided to you for comparison purposes.

1.   Quickly examine these alignments in Jalview to get an overview.

2.   Open HopM1_clustal_10_0.1_codon.fas in MEGA for phylogenetic analysis. You can do this using the file menu or by drag-and-drop. You will need to click through a few prompts:

.    Analyze or Align File = Analyze

.    Input Data Options = Nucleotide Sequences

.    Protein-coding nucleotide sequence data = Yes

.    Select Genetic Code = Standard

3.   Build a Neighbor-Joining (NJ) tree.

.    Use Phylogeny / Construct/Test Neighbor-Joining Tree. You may be asked if you want to use the current data file. Select Yes.

.    Use the following analysis parameters

. Test of Phylogeny = Bootstrap method

. No. of Bootstrap Replications = 100

o Normally I would use 500 or 1000 bootstrap replicates, but you can use

100 for the lab to speed things up.

. Substitution Model = Maximum Composite Likelihood

o If you are doing a protein analysis, I suggest you use Jones-Taylor- Thornton.

. Rates = Uniform

. Gaps/Missing Data Treatment = Pairwise deletion

4.   After you build your tree, a Tree Explorer window will automatically open. This window provides many options for manipulating and viewing your tree. It also provides the option for a caption that provides the details of how the tree was built. The information presented in this caption is the type of information you will want to present in a methods section or figure legend if you publish your results. 

5.   Try using some of the tree manipulation tools available through the menus on the left side of the window. These give you great flexibility is how you view and present your tree. For example, you re-root your tree on a specific branch or midpoint root it. You can collapse and expand nodes. You can change the size of the tree and the fonts.

.    For example, show only bootstrap values greater than or equal to 70

. Statistics/Frequency/Info

. Frequency

. Hide Values Lower Than = 70

.    If you want to rescale your tree to better fit in the window

. Layout

. Auto-size Tree

.    If you want to root the tree on a specific branch

. Subtree

. Root Tree

. Use your mouse to click on the branch and position you want to be the new root

.    To midpoint root your tree

. Layout

. Root on Midpoint

.    Change the layout of the tree to circular or radial format

. Layout

. Tree Style

. Circle or Radiation

6.   Save the tree for later comparison. You can use the File menu to save an analysis session,

which will allow you to go directly back to where you are now, or you can export the tree in Newick format. There are also many other output formats that you might find useful for specific analyses. Finally, you can save the tree image in different formats through the Image menu. For this lab I suggest you Image / Copy to Clipboard and paste into a Word document so you can compare trees from different analyses. Don’t forget to record notes on how you generated  the alignment and tree.

Impact of Substitution Model

We will examine the impact of the substitution model and then use some tools to identify the best model. Remember that substitution models are meant to correct for both biases in the substitution patterns as well as mutational saturation. You can find out more about the substitution models through MEGA help. Wikipedia also has a nice summary of the different models:

https://en.wikipedia.org/wiki/Models  of  DNA_evolution

7.   Generate the nucleotide substitution matrix used for a particular evolutionary model.

.    Use HopM1_clustal_10_0.1.fas

.    Models / Estimate Substitution Matrix (ML).

.    Model/Method = Jukes-Cantor

.    Leave all other parameters to their defaults.

8.   Compare the nucleotide substitution matrix used for the Jukes-Cantor model vs. the Tamura 3- parameter model.

.    Perform. the same analysis as described above but select Tamura 3-Paramter. Note that the results window should stay open from prior analyses so it is easy to compare results.

.    The matrix shows the pairwise probability of substitution. Pay particular attention to the legend below the table, which gives important information, such as the observed nucleotide frequencies for your data.

.    Note, the example substitution matrices shown below were generated from the HopM1_clustal_1_1.fas dataset.

9.   Compare HopM1_clustal_10_0.1.fas Neighbor-Joining trees derived using different substitution models.

.    Use parameters as before, except modify the Model/Method

.    Compare the Maximum Composite Likelihood to Jukes-Cantor to p-distance

.    Look at the scale of the tree and the total branch length (found in the legend).

10. Determine the best evolutionary model. MEGA provides a means to identify the best

evolutionary model based on a maximum likelihood analysis of the data. This is a very useful tool and I recommend using it for selecting the best model.

.    Calculate the best model via Models / Find Best DNA/Protein Models (ML)

.    Examine the NOTE at the bottom of the table. There is a lot of information here,

including information about how to interpret the table and what criteria are used to select the best model.

.    Note, the example analysis shown below was generated from the HopM1_clustal_1_1.fas dataset.

Impact of Phylogenetic Method

Now, let’s examine how using different phylogenetic methods impacts your results.

11. Compare trees.

.    Compare HopM1_clustal_10_0.1_codon.fas trees produced by Neighbor-Joining, Maximum Likelihood, Parsimony, and UPGMA.

. I have provided trees for the latter in Appendix 1.

. The analysis parameters used for each tree are presented in the figure legends.

. Note 1, I reduced the number of bootstrap replicates to 20 simply to speed up the analysis. In reality, I would never use less than 100, and 500 or 1000 would be strongly preferred.

. Note 2, I only show bootstrap scores greater than 50 in each tree. When

comparing the trees, pay particular attention to the clade structure of the tree and the bootstrap support for the nodes.

.    Compare HopM1_clustal_10_0.1_codon.fas trees to HopM1_clustal_1_1.fas trees.

. I have provided trees for the latter in Appendix 2.

. Everything else mentioned above applies here as well.

CASE 2: Species Trees, Gene Trees, and Domain Trees

Let's integrate the skills you have learned over the past few weeks. The goal will be to compare phylogenies produced at three different levels: species, full gene, and conserved domain.

The species phylogeny we will use is the internal transcribed spacer sequence ITS-1. ITS-1 sits between the 18S and 5.8S rRNA genes and is a commonly used molecular marker to identify species. The gene tree will be created from the coding sequence of a gene encoding a receptor-like kinase (RLK). The domain tree will be made from the protein sequence of the conserved kinase domain encoded by each gene.

Species Phylogeny

I have collected ITS-1 sequences (ITS1.fas) from the ITSoneDB (http://itsonedb.cloud.ba.infn.it) for the following species:

.    Arabidopsis thaliana

.    Raphanus sativus (radish)

.    Camelina sativa (camelina, or wild flax)

.    Nicotiana tabacum (tobacco)

.    Solanum lycopersicum (tomato)

.    Phaseolus vulgaris (bean)

.    Glycine max (soybean)

.    Citrus sinensis (orange)

.    Zea mays (corn)

1.   Align the ITS-1 sequences from each of these species (sequences found in the file ITS1.fas) using methods and parameters that you feel will give you the best results

2.   Create a bootstrapped Neighbor-Joining phylogenetic tree from the ITS sequences.

Gene Tree

I have identified homologs of the A. thaliana RLK locus NM_124768 encoding the PSKR2 protein from these species above. Note that I purposefully did not pick the top BLAST hits, so while these are all homologous sequences, many may not be orthologs.

Accession	Species
NM_124768	Arabidopsis thaliana
XM_018635947	Raphanus sativus
XM_010417671	Camelina sativa
XM_016612518	Nicotiana tabacum
NM_001309251	Solanum lycopersicum
XM_007140778	Phaseolus vulgaris
XM_006581242	Glycine max
XM_015530716	Citrus sinensis
XM_008654299	Zea mays

3.   Identify and download the coding sequences of each RLK gene from NCBI.

4.   Align the RLK coding sequences using methods and parameters that you feel will give you the best results.

5.   Create a bootstrapped Neighbor-Joining phylogenetic tree of the RLK genes.

Domain Tree

6.   Identify the conserved kinase domain from the A. thaliana PSKR2 protein. Remember that even if the domain of interest is not listed as a feature in the NCBI accession, you can easily find conserved domains using the NCBI Conserved Domain Database as discussed in step 20 of the NCBI lab. You can directly access this database from any NCBI accession by selecting Identify Conserved Domains from the Analyze this sequence menu on the right side of the page. Don’t  forget that you many have to change the Conserved Domain View from the default Concise Results to the more complete Standard Results.

7.   Based on the domain information from the PSKR2 protein and your MSA, identify the homologous kinase domains from the protein sequences of the eight other species.

8.   Extract the aligned conserved kinase domain.

.    Try to figure out how to block off the appropriate region in Jalview. Hint, once columns are hidden you can save the unhidden alignment to a new file

9.   Realign this domain.

10. Create a bootstrapped Neighbor-Joining phylogenetic tree.

Comparative Analysis

11. Finally, compare the ITS-1 species tree, the PSKR2 gene tree, and the PSKR2 kinase domain tree. This comparison is your quiz.

Your quiz will be a summary of the analysis results from CASE 2: Species Trees, Gene Trees, and Domain Trees. Record your results in a file (Word or pdf) and upload it to Lab 6 Quiz: Phylogenetics 2. Minimally, you should show the following: .    The phylogenetic analyses: o ITS-1 species tree o PSKR2 gene tree o PSKR2 kinase domain tree .    Summary information on how you performed each analysis. Enough information should be presented to recreate your work. In addition to explaining the approach and choice of parameters for the phylogenetic analysis, don't forget to explain your alignment and model selection. .    Very short discussion / analysis of the similarities and differences between these trees and explanation. .    Keep this report tightly focused on what analyses you did, why you selected the analyses and parameters, and a basic comparative analysis of the results.