Phylogenetic methods - computer practicals
Parsimony
Maximum likelihood
Bayesian inference
Species tree
home
Department of Botany, Charles University, Prague (2021-2024)
Maximum likelihood
Basic comments to RAxML are here.
Alternativelly, you can use RAxML-NG here.
Training dataset is here. It consists of
ITS and matK alignments of eight samples from the ginger family (Zingiberaceae) published by
de Boer et al. 2018. Outgroup taxon is Siphonochilus kirkii.
Moreover, there is an AFLP (binary) matrix based on the dataset of Cardamine
impatiens.A ZIP file with necessary software is
here.
Download it and unzip.
RAxML - Randomized Axelerated Maximum Likelihood
FigTree
- phylogenetic tree viewer
Other software recommended to install
Tasks 1 (to work
with Cardamine dataset)
after you are done submit the answers using
this Google form
If an advice is required, contact Tomá Fér: tomas.fer_AT_natur.cuni.cz
1. Create the maximum likelihood tree of the
Cardamine AFLP dataset
-
put the matrix in FASTA format to the same folder
where is the RAxML binary
-
navigate to that folder in Total Commander and click
on Commands - Open command prompt window
-
type this command and press Enter
raxmlHPC -m BINGAMMA -p 12345 -s impatiens_01.fas
-# 5 -n resultBIN
Question 1: What is the likelihood of the best
tree?
-
open the resulting tree (RAxML_bestTree.resultBIN) in
FigTree and display as unrooted
Question 2: How many groups of individuals do you
see?
Tasks 2 (to work
with Amomum dataset)
- Create best ML tree for the ITS
dataset using GTRGAMMA model
- run this command
raxmlHPC -m GTRGAMMA -p 12345 -s
Amomum_ITS.trimmed.fas -# 10 -n resultDNA
Question 3: What is the difference in likelihood
between the tree with highest and the smallest likelihood?
- open all 10 resulting trees in FigTree (*.RUN0 -
*.RUN9), root trees with outgroup (Siphonochilus kirkii)
Question 4: What are the differences among trees?
- Create the best ML tree for the
matK dataset
- repeat the previous steps for the
Amomum_matK.trimmed.fas (change the parameter -n
to bestML)
- open best tree (RAxML_bestTree.*) in FigTree
Question 5: What is
the difference among the best ITS and the best
matK tree?
- Run the rapid bootstrap for the
matK dataset
- run the full analysis using this command
raxmlHPC -f a -m GTRGAMMA -p 12345 -x 12345 -#
100 -s Amomum_matK.trimmed.fas -n rbs
- open the resulting tree (RAxML_bipartitions.rbs) in
FigTree
Question 6: What is the lowest bootstrap support
value?
- run the rapid bootstrap analysis again (but change
the -n parameter to, e.g., rbs2, and
-p and -x
to, e.g., 12346)
Question 7: Are all the bootstrap values the
same? If not, why?
- repeat the analysis twice again but now with 1,000
bootstrap replicates (-# 1000), always
changing -p, -x
and -n
- Run the standard bootstrap
analysis for the matK dataset
- the best ML tree was already created in the step 3
- create 100 standard bootstrap replicate trees
raxmlHPC -m GTRGAMMA -p 12345 -b 12345 -# 100 -s
Amomum_matK.trimmed.fas -n boot
- map bootstrap values onto the best ML tree
raxmlHPC -m GTRGAMMA -p 12345 -f b -t
RAxML_bestTree.bestML -z RAxML_bootstrap.boot -n finalboot
Question 8: Are the standard bootstrap values
same as rapid bootstrap values? Is there any trend?
- Create the rapid and standard
bootstrap analysis for the concatenated and partitioned dataset
- look at the file Amomum_partitions.trimmed.txt how
the partitions are specified
- run similar analysis as in steps 3-5 but specify
the partition file using parameter -q (and
do not forget to specify different name for parameter
-n)
raxmlHPC -m GTRGAMMA -p 12345 -s
Amomum_concatenated.trimmed.fas -q Amomum_partitions.trimmed.txt -# 10 -n
concatbestML
- continue running bootstrap replicates and mapping
to the bestML tree (similar to step 5), again, do not forget
-q
- do the rapid bootstrap analysis (similar to step
4), specify -q
Question 9: Is the topology of ITS, matK
and concatenated tree the same or what are the differences? (Always compare
rooted trees)
Question 10: Compare standard and rapid bootstrap
values for the concatenated dataset.
Thank you for participating in RAxML practicals...