Skip to content
A wrapper for YASRA and associated parsing scripts
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.gitignore
INSTALL.txt
LICENSE.txt
README.md
alignreads.py
default_configuration.py
install.py
makeconsensus.py
readtools.py
readtools_exceptions.py
runyasra.py

README.md

Alignreads

Alignreads is a wrapper for YASRA (http://www.bx.psu.edu/miller_lab/). The principal function of alignreads is to facilitate easy execution of YASRA and to parse its output.

YASRA is a reference guided assembler that has the ability to extend the edges of alignments de novo interactively. The minimum inputs are a reference sequence and reads to be aligned, but there are many options. Use alignreads -h after installation to see a full list of options.

Alignreads has the following dependencies:

  • YASRA
  • lastz (used by YASRA)
  • nucmer (from the MUMmer tool kit)
  • Biopython (used to parse YASRA's output)

The installer of alignreads can typically download and install most of these automatically. See INSTALL.txt for details.

YASRA was created by Aakrosh Ratan at Pennsylvania State University in the Miller lab.

Alignreads was created by Zachary Foster at Oregon State University in the Liston lab.

Installation

Alignreads has an installation script that attempts to download and install all dependencies where possible. This automatic installation should work on most system configurations with access to the internet and common compiling software (e.g. gcc)

In the examples below, replace terms between < and > with values appropriate to your system and needs:

Downloading alignreads

Alignreads releases are available from Git Hub at https://github.com/zachary-foster/alignreads/releases. The wget Linux command can be used to download a release from the command line:

wget https://github.com/zachary-foster/alignreads/releases/<version to install>

You can also download the current development version of alignreads using git:

git clone https://github.com/zachary-foster/alignreads

Unpacking the release

Use tar to unpack the alignreads directory:

tar -xvf <version to install>.tar.gz
cd <version to install>.tar.gz

Installer help information

To see what options are available for the installation process, run the installation script without any arguments:

python install.py

Running the interactive installer

By default, the installer allows you to interactively choose which version of dependencies to install and attempts to download and install them:

python install.py <path to installation location>

The <path to installation location> will be the where alignreads will be installed. A folder named alignreads will be created there.

Non-interactive installation

Alternatively, to automatically install the recommended versions of dependencies non-interactively, the -i / --not-interactive option can be used:

python install.py -i <path to installation location>

Using currently installed dependecies

If some of the dependencies are already installed on you system, you can use those versions by including the corresponding installer option (See installer help menu). If the already-installed dependencies are in your PATH (i.e. you can use them from the command line in any directory), just supply the corresponding option without arguments. For example, if lastz is already installed and in your PATH:

python install.py --lastz <path to installation location>

If a dependency is installed, but not in PATH, then give the corresponding option the location to find it:

python install.py --lastz <path to lastz> <path to installation location>

Putting alignreads in your shell's search path

The installation of alignreads contains a executable shell script that simply executes alignreads with a specified instance of python. This shell script allows for alignreads to be run by typing alignreads instead of python alignreads.py. It is often convenient to put this executable in your shell's search path, so that typing alignreads runs the program regardless of current working directory.

How this is done depends on what shell you are using. To find out what shell you are using, paste the following command into your terminal:

ps -p $$ -o cmd=

For Bash/Sh/Ksh:

In your home directory, there should be a .bashrc or .profile file depending on your system. This file contains a list of bash commands that are run each time bash is run (i.e. a terminal is opened). To put alignreads in your search path permanently, add something like the following line to the end of that file.

export PATH="$PATH:<path to installation>"

Here is an example of the line I would add to .bashrc if I installed alignreads in /home/fosterz on my system.

export PATH="$PATH:/home/fosterz/alignreads"

For Csh/Tcsh:

In your home directory, there should be a .cshrc or .login file, depending on your system. This file contains a list of csh commands that are run each time csh is run. To put alignreads in your search path permanently, add something like the following line to the end of that file.

setenv PATH $PATH:<path to installation>

Here is an example of the line I would add to .cshrc if I installed alignreads in /home/fosterz on my system.

setenv PATH $PATH:/home/fosterz/alignreads

Adding to PATH without a text editor

If you want a one-line solution or dont want to use a command line text editor to make your changes to the shell configuration file, you can cd into the directory you want to add to your shell's search path and type:

echo "export PATH=$(pwd):\${PATH}" >> ~/.bashrc

This example assumes you are using bash, but something similar can be used for other shells as well.

Running alignreads

Alignreads has an intimidating amount of arguments because it gives access to the arguments of its constituent programs. However, most of these arguments are rarely used. Running alignreads without arguments will print a help menu with all the options and their default values.

The simplest alignreads command possible uses default values for all of the options. For example, to assemble reads in a fasta file named read.fa using a reference in a file called reference.fa, assuming both files are in the current working directory and alignreads is in your shell's search path, type:

alignreads reads.fa reference.fa

In most cases, at least some options should be used. Examples and explanations of commonly needed options follow.

General alignreads options

-y / --output-directory

Specify the location of the output directory. By default, alignreads creates an output directory for each run in the current working directory.

For example, to create the output directory in the folder ~/my_assemblies regardless of current working directory:

alignreads reads.fa reference.fa --output-directory ~/my_assemblies

-c / --config-file

Alignreads uses a configuration file to specify the default values of all options. The configuration file is simply a python script that defines variables. The default configuration file is located in the installed alignreads folder and is called default_configuration.py. You can make copies of this file and change the values to specify your own defaults for different tasks.

For example, to use a configuration file called bacterial_configuration.py that specifies the default option values for assembling bacterial genomes:

alignreads reads.fa reference.fa --config-file bacterial_configuration.py

Any other options used on the command line will override the defaults in the configuration file if used together.

For example, say you are only assembling a specific part of a bacterial genome (i.e. not circular). You could run:

alignreads reads.fa reference.fa --config-file bacterial_configuration.py -o linear

Contig assembly options

-a / --single-step

YASRA uses an iterative process to improve assemblies. If additional iterations after the first do not improve the assembly YASRA throws an error. If this happens, you have to use the -a / --single-step option to prevent the error. A future goal is to circumvent this inconvenient behavior.

For example:

alignreads reads.fa reference.fa --single-step

-t / --read-type

The type of reads, either 454 or solexa.

  • 454: longer, less accurate Roche 454 pyrosequencer reads.
  • solexa: shorter, more accurate Illumina reads.

For example, if the reads being used were produced by an Illumina HiSeq:

alignreads reads.fa reference.fa -t solexa

-o / --read-orientation

Whether the thing being assembled is a linear or circular DNA molecule.

For example, if a chromosome is being assembled:

alignreads reads.fa reference.fa -o linear

-p / --percent-identity

How similar the reference is to the sequence being assembled. The interpretation of this option is effected by --read-type. Possible values are listed below.

For 454:

  • same : about 98% identity between target & reference
  • high : about 95% identity between target & reference
  • `medium : about 90% identity between target & reference
  • low : about 85% identity between target & reference
  • verylow : about 75% identity between target & reference
  • desperate : really low identity (rather slow)

For Solexa/Illumina:

  • same : about 95% identity between target & reference
  • medium : about 85% identity between target & reference
  • desperate : low scores (rather slow)

For example, to assemble Illumina reads with a closely related reference:

alignreads reads.fa reference.fa -t solexa -p same

Contig alignment options

The program nucmer from the MUMmer tool kit is used to align contigs produced by YASRA to the reference. nucmer has many options that control this process. To fully understand the effects of the options, a understanding of how nucmer works is needed. I recommend looking over the nucmer documentation (http://mummer.sourceforge.net/manual/#nucmer).

Consensus generation options

Alignreads makes a consensus of all of the aligned contigs. There are a few options that are almost essential for this process to work correctly. All of the following options accept a range of numbers in the format <min>-<max> (e.g. 5-10), inclusive. The or the can be left out to indicate no minimum/maximum limit (e.g. 5- means "at least five"). Multiple ranges can be specified if separated by commas (e.g. 1-3,100- means "one to three or over 100").

-g / --quality-read-filter

The acceptable quality scores for reads. Reads outside this range will not be included in the consensus sequences.

-m / --depth-position-masking

The range(s) of acceptable read depth. Positions outside this range will be masked in the consensus sequences.

For example, to mask all positions with less than 10 read depth in the consensus sequence:

alignreads reads.fa reference.fa -m 10-

-b / --proportion-base-filter

The acceptable range(s) for the proportion of bases at a given position that support a given call. Nucleotides with proportions outside of this range will be ignored when condensing the position to an IUPAC character in the consensus.

For example, to ignore all SNPs that have a less than 10% frequency:

alignreads reads.fa reference.fa -b 0.1-

-d / --depth-position-filter

Set the depth range(s) for position filtering. Positions outside this range will not be included in the consensus sequences.

For example, to remove indels that occur less than 5% of the time:

alignreads reads.fa reference.fa -d 0.05-

Making multiple consensus sequences per assembly

Running YASRA can take a long time whereas aligning the contigs and making a consensus sequence is relatively quick. This means that a lot of time can be wasted optimizing contig alignment and consensus generation parameters if a new assembly is done each try. Therefore, it is possible to redo everything downstream of the assembly step with different options by supplying alignreads with the path to a previously-generated output directory instead of the read and reference file. A new sub-directory will be created within the old output directory for each new run.

For example, to make a new consensus from a previous assembly stored in alignreads output directory named 454.fa_rhino_template.fa_Wed-Dec-17-15:36:23-2014:

alignreads 454.fa_rhino_template.fa_Wed-Dec-17-15:36:23-2014 -d 0.15-

Alignreads output

Alignreads produces an output directory with a name composed of the names of the read and reference file and a time stamp. This is created in current working directory unless the -y / --output-directory option is used. The output directory contains a sub-directory YASRA_related_files that contains files produced by and associated with YASRA. The output directory also contains at least one directory called alignment_# that contains all the post-assembly files. Multiple alignment_# sub-directories can exist of the post-assembly analysis is run multiple times.

YASRA_related_files contents:

  • alignments_reads.fa_reference.fa.sam : SAM file of how the reads align to each contig
  • Command_Line_Record.txt : The command line used to produce the assembly
  • contigs_reads.fa_reference.fa.ace : ACE file containing read and contig information
  • Final_Assembly_reads.fa_reference.fa : FASTA file of contigs
  • Makefile : Used to run YASRA
  • link to read file
  • link to reference file

alignment_# contents:

  • alignments_reads.fa_reference.fa.sam.fa : FASTA file of aligned contigs
  • alignments_reads.fa_reference.fa.sam.fa_aligned.fa : FASTA file of aligned consensus sequence, reference, and contigs
  • alignments_454.fa_rhino_template.fa.sam.fa.delta : File produced by nucmer indicating alignments
  • Execution_info.txt : Info on parameters used and run-time log
  • nucmer_standard_error.txt : Nucmer standard error output.
  • nucmer_standard_output.txt : Nucmer standard out output.
You can’t perform that action at this time.