- Installation using
vadr-install.sh - Setting environment variables
- Verifying successful installation
- VADR-specific modifications to FASTA
- Further information
The file vadr-install.sh is an executable file for downloading
and/or installing VADR and its dependencies. That file is located
online at github. To install the latest version of VADR download this
file:
https://raw.githubusercontent.com/ncbi/vadr/master/vadr-install.sh
To download any specific release/version, for example version 1.0 download
the corresponding vadr-install.sh file for that release/version
(prior to version 1.0, the name of the installation script was
install.sh, not vadr-install.sh):
https://raw.githubusercontent.com/ncbi/vadr/vadr-1.0/vadr-install.sh
Copy the vadr-install.sh file into the directory in which you want
to install VADR. A good name for that directory is
vadr-install-dir. Then move into that directory and run one of the
following three commands depending on whether you are installing on a
Linux, Mac/OSX system with Apple silicon chips (M1/M2) or Mac/OSX
system with intel chips:
sh ./vadr-install.sh linux
OR
sh ./vadr-install.sh macosx-silicon
OR
sh ./vadr-install.sh macosx-intel
The vadr-install.sh command will create several directories in the
current directory. It will download and install VADR and the required
module libraries sequip,
Bio-Easel, as well as
Infernal, the NCBI BLAST package,
William Pearson's FASTA package (for either Linux or Mac/OSX), and
Heng Li's (minimap2)[https://github.com/lh3/minimap2]. (The VADR
installation script modifies the FASTA package slightly prior to
installing it to allow alignment of long sequences up to 250Kb using
the glsearch executable. This is described more here.)
The installation requires that you have the perl Inline module installed on your system. If not, the installation script may fail. If this happens, read this.
When vadr-install.sh is finished running it will print important
instructions to the screen that explain how to modify your environment
variables so that you can run the VADR scripts, as discussed next.
Installation on a mac will only work if Xcode has been installed. Xcode is available for download from the Mac app store.
Alternatively, you can install VADR in two steps, downloading the dependencies and source code files in step 1, and building packages in step 2. This alternative approach, using the following two commands, may be more convenient for some users. The build step must be run from the same directory that the download step was run in:
sh ./vadr-install.sh linux download
sh ./vadr-install.sh linux build
These commands should have macosx-silicon or macosx-intel instead
of linux if you are installing on Mac/OSX.
If installation or do-install-tests-local.sh fails because the Inline, LWP or Mozilla::CA perl modules are not installed...
The perl Inline::C, LWP::Simple, LWP::Protocol::https and Mozilla::CA modules
must be installed prior to installation. (This
page
lists three ways to check if a perl module is installed.)
You can install these modules using cpan with these commands:
cpan install Inline::C
cpan install LWP::Simple
cpan install LWP::Protocol::https
cpan install Mozilla::CA
However, for Mac/OSX these commands may not work if you have not installed the "Command line tools for Xcode" or "XCode" packages. You can download Xcode from the Mac App Store for free.
It is possible that you have Inline installed but not for C, if this is the case, you might see an error message like this:
Error. You have specified 'C' as an Inline programming language.
I currently only know about the following languages:
Foo, foo
If this is the case, try installing Inline::C with this command:
cpan install Inline::C
As mentioned above, when you run vadr-install.sh, instructions will be
output about how to change your environment variables so that you can
run the VADR scripts. Those instructions are also included below for
reference, but without the actual path to where you ran vadr-install.sh
(below it is replaced with <full path to directory in which you ran vadr-install.sh>)
********************************************************
The final step is to update your environment variables.
(See https://github.com/ncbi/vadr/blob/1.6.3/documentation/install.md for more information.)
If you are using the bash or zsh shell (zsh is default in MacOS/X as
of v10.15 (Catalina)), add the following lines to the end of your
'.bashrc' or '.zshrc' file in your home directory:
export VADRINSTALLDIR=<full path to directory in which you ran vadr-install.sh>
export VADRSCRIPTSDIR="$VADRINSTALLDIR/vadr"
export VADRMODELDIR="$VADRINSTALLDIR/vadr-models-calici"
export VADRINFERNALDIR="$VADRINSTALLDIR/infernal/binaries"
export VADREASELDIR="$VADRINSTALLDIR/infernal/binaries"
export VADRHMMERDIR="$VADRINSTALLDIR/infernal/binaries"
export VADRBIOEASELDIR="$VADRINSTALLDIR/Bio-Easel"
export VADRSEQUIPDIR="$VADRINSTALLDIR/sequip"
export VADRBLASTDIR="$VADRINSTALLDIR/ncbi-blast/bin"
export VADRFASTADIR="$VADRINSTALLDIR/fasta/bin"
export VADRMINIMAP2DIR="$VADRINSTALLDIR/minimap2"
export PERL5LIB="$VADRSCRIPTSDIR":"$VADRSEQUIPDIR":"$VADRBIOEASELDIR/blib/lib":"$VADRBIOEASELDIR/blib/arch":"$PERL5LIB"
export PATH="$VADRSCRIPTSDIR":"$PATH"
After adding the export lines to your .bashrc or .zshrc file, source that file
to update your current environment with the command:
source ~/.bashrc
OR
source ~/.zshrc
---
If you are using the C shell, add the following
lines to the end of your '.cshrc' file in your home
directory:
setenv VADRINSTALLDIR "<full path to directory in which you ran vadr-install.sh>"
setenv VADRSCRIPTSDIR "$VADRINSTALLDIR/vadr"
setenv VADRMODELDIR "$VADRINSTALLDIR/vadr-models-calici"
setenv VADRINFERNALDIR "$VADRINSTALLDIR/infernal/binaries"
setenv VADRHMMERDIR "$VADRINSTALLDIR/infernal/binaries"
setenv VADREASELDIR "$VADRINSTALLDIR/infernal/binaries"
setenv VADRBIOEASELDIR "$VADRINSTALLDIR/Bio-Easel"
setenv VADRSEQUIPDIR "$VADRINSTALLDIR/sequip"
setenv VADRBLASTDIR "$VADRINSTALLDIR/ncbi-blast/bin"
setenv VADRFASTADIR "$VADRINSTALLDIR/fasta/bin"
setenv VADRMINIMAP2DIR "$VADRINSTALLDIR/minimap2"
setenv PERL5LIB "$VADRSCRIPTSDIR":"$VADRSEQUIPDIR":"$VADRBIOEASELDIR/blib/lib":"$VADRBIOEASELDIR/blib/arch":"$PERL5LIB"
setenv PATH "$VADRSCRIPTSDIR":"$PATH"
After adding the setenv lines to your .cshrc file, source that file
to update your current environment with the command:
source ~/.cshrc
(To determine which shell you use, type: 'echo $SHELL')
********************************************************
If you use bash or zsh, change the PERL5LIB line in your ~/.bashrc or
~/.zshrc file to:
export PERL5LIB="$VADRSCRIPTSDIR":"$VADRSEQUIPDIR":"$VADRBIOEASELDIR/blib/lib":"$VADRBIOEASELDIR/blib/arch"
or if you use C shell, change the PERL5LIB line in your ~/.cshrc
file to:
setenv PERL5LIB "$VADRSCRIPTSDIR":"$VADRSEQUIPDIR":"$VADRBIOEASELDIR/blib/lib":"$VADRBIOEASELDIR/blib/arch"
And then execute source ~/.bashrc, source ~/.zshrc, or source ~/.cshrc again.
The VADR package includes some tests you can run to make sure that your installation was successful and that your environment variables are set-up correctly.
These are several shell scripts for running tests; with respect to the
installation directory they are in the directory vadr/testfiles/ and
start with do- and end with .sh.
At a minimum, you should run the
vadr/testfiles/do-install-tests-local.sh script to make sure VADR installed
correctly. They should pass, as shown below.
The script vadr/testfiles/do-all-tests.sh will run all tests, but be
warned that script may take up to an hour or so to run. All the tests
should pass.
To run all tests, execute:
$VADRSCRIPTSDIR/testfiles/do-all-tests.sh
To run only the do-install-tests-local.sh script, execute:
$VADRSCRIPTSDIR/testfiles/do-install-tests-local.sh
This scripts can take up to several minutes to run. If something goes wrong, the script will likely exit quickly.
If the script fails immediately with a warning like:
Can't locate LWP/Simple.pm in @INC (you may need to install the LWP::Simple module)
Or something similar but with Inline or Mozilla::CA instead of
LWP, then you will need to install the perl LWP, Mozilla::CA and/or Inline
modules as described here.
Below is an example of the expected output for
do-install-tests-local.sh:
# v-test.pl :: test VADR scripts [TEST SCRIPT]
# VADR 1.6.2 (Dec 2023)
# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
# date: Mon Dec 4 10:29:11 2023
#
# test file: /net/intdev/oblast01/infernal/notebook/23_0925_vadr_1p6_and_1p6p1_and_1p6p2_releases/test-install13/vadr/testfiles/noro.r10.local.testin
# output directory: vt-n10-local
# forcing directory overwrite: yes [-f]
# if output files listed in testin file already exist, remove them: yes [--rmout]
# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
# Parsing test file ... done. [ 0.0 seconds]
##teamcity[testStarted name='annotate-noro-10-local' captureStandardOutput='true']
# Running command 1 [annotate-noro-10-local] ... done. [ 35.4 seconds]
# checking va-noro.r10/va-noro.r10.vadr.pass.tbl ... pass
# checking va-noro.r10/va-noro.r10.vadr.fail.tbl ... pass
# checking va-noro.r10/va-noro.r10.vadr.sqa ... pass
# checking va-noro.r10/va-noro.r10.vadr.sqc ... pass
# checking va-noro.r10/va-noro.r10.vadr.ftr ... pass
# checking va-noro.r10/va-noro.r10.vadr.sgm ... pass
# checking va-noro.r10/va-noro.r10.vadr.mdl ... pass
# checking va-noro.r10/va-noro.r10.vadr.alt ... pass
# checking va-noro.r10/va-noro.r10.vadr.alc ... pass
# removing directory va-noro.r10 ... done
##teamcity[testFinished name='annotate-noro-10-local']
#
#
# PASS: all 9 files were created correctly.
#
# Output printed to screen saved in: vt-n10-local.vadr.log
# List of executed commands saved in: vt-n10-local.vadr.cmd
# List and description of all output files saved in: vt-n10-local.vadr.list
#
# All output files created in directory ./vt-n10-local/
#
# Elapsed time: 00:00:36.05
# hh:mm:ss
#
[ok]
# v-test.pl :: test VADR scripts [TEST SCRIPT]
# VADR 1.6.2 (Dec 2023)
# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
# date: Mon Dec 4 10:29:48 2023
#
# test file: /net/intdev/oblast01/infernal/notebook/23_0925_vadr_1p6_and_1p6p1_and_1p6p2_releases/test-install13/vadr/testfiles/dengue.r5.local.testin
# output directory: vt-d5-local
# forcing directory overwrite: yes [-f]
# if output files listed in testin file already exist, remove them: yes [--rmout]
# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
# Parsing test file ... done. [ 0.0 seconds]
##teamcity[testStarted name='annotate-dengue-5-local' captureStandardOutput='true']
# Running command 1 [annotate-dengue-5-local] ... done. [ 78.5 seconds]
# checking va-dengue.r5/va-dengue.r5.vadr.pass.tbl ... pass
# checking va-dengue.r5/va-dengue.r5.vadr.fail.tbl ... pass
# checking va-dengue.r5/va-dengue.r5.vadr.sqa ... pass
# checking va-dengue.r5/va-dengue.r5.vadr.sqc ... pass
# checking va-dengue.r5/va-dengue.r5.vadr.ftr ... pass
# checking va-dengue.r5/va-dengue.r5.vadr.sgm ... pass
# checking va-dengue.r5/va-dengue.r5.vadr.mdl ... pass
# checking va-dengue.r5/va-dengue.r5.vadr.alt ... pass
# checking va-dengue.r5/va-dengue.r5.vadr.alc ... pass
# removing directory va-dengue.r5 ... done
##teamcity[testFinished name='annotate-dengue-5-local']
#
#
# PASS: all 9 files were created correctly.
#
# Output printed to screen saved in: vt-d5-local.vadr.log
# List of executed commands saved in: vt-d5-local.vadr.cmd
# List and description of all output files saved in: vt-d5-local.vadr.list
#
# All output files created in directory ./vt-d5-local/
#
# Elapsed time: 00:01:19.16
# hh:mm:ss
#
[ok]
Success: all tests passed [do-install-tests-local.sh]
The most important line is the final line:
Success: all tests passed [do-install-tests-local.sh]
This means that the test has passed. You should see similar lines if you run the other tests. If you do not and need help figuring out why, email me at [email protected].
To enable alignment of long sequences (up to 250Kb) with the
glsearch program of FASTA, some changes are made to the src/defs.h
file by the vadr-install.sh script prior to compilation of
FASTA. These changes can be seen in the patch file
vadr/fasta-mods/vadr-fasta-defs.patch
patch file. Additionally, to
avoid possible numerical overflow issues with glsearch scores when
aligning long sequences, the FASTA Makefile is modified so that the
non-vectorized version of glsearch is used. These changes can be
seen in the patch files
vadr/fasta-mods/vadr-fasta-Makefile.linux.patch and
vadr/fasta-mods/vadr-fasta-Makefile.os_x86_64.patch. See the
vadr/fasta-mods/README.txt for more information on how these patch
files were created.
When aligning very long sequences (more than 30Kb), it is recommended to not
build a CM with v-build.pl (i.e. use the --skipbuild option),
and it is recommended to use the -s and --glsearch options with v-annotate.pl.
v-build.plexample usage and command-line optionsv-annotate.plexample usage, command-line options and alert information- Explanations and examples of
v-annotate.pldetailed alert and error messages - VADR output file formats
- Available VADR model files (github wiki)
- SARS-CoV-2 annotation (github wiki)
- Development notes and instructions (github wiki)