bch441-work-abc-units/BIN-PHYLO-Tree_building.R

143 lines
5.2 KiB
R
Raw Normal View History

2017-11-01 13:44:24 +00:00
# BIN-PHYLO-Tree_building.R
2017-09-12 20:09:20 +00:00
#
# Purpose: A Bioinformatics Course:
2017-11-01 13:44:24 +00:00
# R code accompanying the BIN-PHYLO-Tree_building unit.
2017-09-12 20:09:20 +00:00
#
2017-11-01 13:44:24 +00:00
# Version: 1.0
2017-09-12 20:09:20 +00:00
#
2017-11-01 13:44:24 +00:00
# Date: 2017 10. 31
2017-09-12 20:09:20 +00:00
# Author: Boris Steipe (boris.steipe@utoronto.ca)
#
# Versions:
2017-11-01 13:44:24 +00:00
# 1.0 First 2017 version
2017-09-12 20:09:20 +00:00
# 0.1 First code copied from 2016 material.
2017-11-01 13:44:24 +00:00
#
2017-09-12 20:09:20 +00:00
#
# TODO:
2017-11-02 03:30:01 +00:00
# Add MrBayes
# https://cran.r-project.org/web/packages/phangorn/vignettes/IntertwiningTreesAndNetworks.html
2017-09-12 20:09:20 +00:00
#
# == DO NOT SIMPLY source() THIS FILE! =======================================
2017-11-01 13:44:24 +00:00
#
2017-09-12 20:09:20 +00:00
# If there are portions you don't understand, use R's help system, Google for an
# answer, or ask your instructor. Don't continue if you don't understand what's
# going on. That's not how it works ...
2017-11-01 13:44:24 +00:00
#
2017-09-12 20:09:20 +00:00
# ==============================================================================
2017-11-02 03:30:01 +00:00
#TOC> ==========================================================================
2017-11-23 12:19:31 +00:00
#TOC>
2017-11-02 03:30:01 +00:00
#TOC> Section Title Line
#TOC> -------------------------------------------------------
#TOC> 1 Calculating Trees 43
#TOC> 1.1 PROMLPATH ... 64
#TOC> 1.1.1 ... on the Mac 69
#TOC> 1.1.2 ... on Windows 80
#TOC> 1.1.3 ... on Linux 94
#TOC> 1.1.4 Confirming PROMLPATH 99
2017-11-23 12:19:31 +00:00
#TOC> 1.2 Building a maximum likelihood tree 108
#TOC>
2017-11-02 03:30:01 +00:00
#TOC> ==========================================================================
2017-11-01 13:44:24 +00:00
2017-11-02 03:30:01 +00:00
# = 1 Calculating Trees ===================================================
2017-11-01 13:44:24 +00:00
2017-09-12 20:09:20 +00:00
# Follow the instructions found at phylip's home on the Web to install. If you
# are on a Windows computer, take note of the installation directory.
# After you have installed Phylip on your computer, install the R package that
# provides an interface to the Phylip functions.
if (!require(Rphylip, quietly=TRUE)) {
install.packages("Rphylip")
library(Rphylip)
}
# Package information:
# library(help = Rphylip) # basic information
# browseVignettes("Rphylip") # available vignettes
# data(package = "Rphylip") # available datasets
2017-09-12 20:09:20 +00:00
# This will install RPhylip, as well as its dependency, the package "ape".
2017-11-02 03:30:01 +00:00
# == 1.1 PROMLPATH ... =====================================================
2017-09-12 20:09:20 +00:00
# The next part may be tricky. You will need to figure out where
# on your computer Phylip has been installed and define the path
# to the proml program that calculates a maximum-likelihood tree.
2017-11-23 12:19:31 +00:00
# === 1.1.1 ... on the Mac
2017-09-12 20:09:20 +00:00
# On the Mac, the standard installation places a phylip folder
# in the /Applications directory. That folder contains all the
# individual phylip programs as <name>.app files. These are not
# the actual executables, but "app" files are actually directories
# that contain the required resources for a program to run.
# The executable is in a subdirectory and you can point Rphylip
# directly to that subdirectory to find the program it needs:
# PROMLPATH <- "/Applications/phylip-3.695/exe/proml.app/Contents/MacOS"
2017-11-23 12:19:31 +00:00
# === 1.1.2 ... on Windows
# On Windows you need to know where the programs have been installed, and you
2017-09-12 20:09:20 +00:00
# need to specify a path that is correct for the Windows OS. Find the folder
# that is named "exe", and right-click to inspect its properties. The path
# should be listed among them.
# If the path looks like "C:\Users\Meng\Programs\phylip-3.695\exe", then your
# assignment has to be
# PROMLPATH <- "C:/Users/Meng/Programs/phylip-3.695/exe"
# (Note: "/", not "\")
# I have heard that your path must not contain spaces, and it is prudent to
# avoid other special characters as well.
2017-11-23 12:19:31 +00:00
# === 1.1.3 ... on Linux
2017-09-12 20:09:20 +00:00
# If you are running Linux I trust you know what to do. It's probably
# something like
# PROMLPATH <- "/usr/local/phylip-3.695/bin"
2017-11-23 12:19:31 +00:00
# === 1.1.4 Confirming PROMLPATH
2017-09-12 20:09:20 +00:00
# Confirm that the settings are right.
PROMLPATH # returns the path
list.dirs(PROMLPATH) # returns the directories in that path
2017-11-01 13:44:24 +00:00
list.files(PROMLPATH) # lists the files [1] "proml" "proml.command"
2017-09-12 20:09:20 +00:00
# If "proml" is NOT among the files that the last command returns, you
2017-11-01 13:44:24 +00:00
# can't continue. Ask on the mailing list for advice.
2017-09-12 20:09:20 +00:00
2017-11-23 12:19:31 +00:00
# == 1.2 Building a maximum likelihood tree ================================
2017-11-01 13:44:24 +00:00
# Now read the mfa file you have saved in the BIB-PHYLO-Data_preparation unit,
# as a "proseq" object with the read.protein() function of the RPhylip package:
2017-09-12 20:09:20 +00:00
2017-11-01 13:44:24 +00:00
apsIn <- read.protein("APSESphyloSet.mfa")
2017-09-12 20:09:20 +00:00
# ... and you are ready to build a tree.
2017-11-02 03:30:01 +00:00
# There are many fast options in PHYLIP - we will use the most _accurate_ one
# that it has: proml, a maximum-likelihood tree building program for protein
# data.
2017-09-12 20:09:20 +00:00
# Building maximum-likelihood trees can eat as much computer time
# as you can throw at it. Calculating a tree of 48 APSES domains
# with default parameters of Rproml() runs for more than half a day
# on my computer. But we have only twelve sequences here, so the
2017-11-02 03:30:01 +00:00
# process will take us about 5 to 10 minutes. Run this, and anjoy a good cup
# of coffee while you are waiting.
2017-09-12 20:09:20 +00:00
apsTree <- Rproml(apsIn, path=PROMLPATH)
# A quick first look:
plot(apsTree)
2017-11-01 13:44:24 +00:00
# save your tree:
save(apsTree, file = "APSEStreeRproml.RData")
2017-09-12 20:09:20 +00:00
2017-11-01 13:44:24 +00:00
# If this did not work, ask for advice.
2017-09-12 20:09:20 +00:00
# [END]