Personal tools
Home R├Ątsch Lab Supplements SAFT

SAFT: Simple Alignment Filtering Tool

News

Jul. 10, 2012:SAFT has been moved to the RNA-Geeq Toolbox. The download will be available from August 1st, 2012
Jul. 07, 2011:Minor Bug fixes, examples added. Download
Jun. 08, 2011:An updated Version of SAFT is now available.
Apr. 15, 2011:The first version of the Simple Alignment Filter Toolbox (SAFT) has been released.

Information about SAFT

SAFT is the abbreviation for Simple Alignment Filtering Tool and was developed by Andre Kahles at the Friedrich Miescher Laboratory of the Max Planck Society in 2010 and 2011.

Currently latest Version: 0.2 (alpha) - 2011.06.08

Installation and Dependencies

BioPython
For the correct parsing of GFF3 file SAFT relies on BioPython. BioPython is available at www.biopython.org/ Please make sure, that BioPython is available in your PYTHONPATH.
SAMtools
If you want to use BAM files as input, you need a working version of SAMtools (>= 0.1.7a) available in your PATH. You can download SAMtools from http://samtools.sourceforge.net/
SAFT
To install SAFT, download the latest archive from http://raetschlab.org/suppl/srm-eval/SAFT . Now, move the archive to a place of choice and unpack it. The directory SAFT-{VERSION} will be automatically created. You are ready to run SAFT!

Description of the single tools

Following tools are available:

  1. a - filter_alignment.py
  1. b - filter_features.py
  2. a - gen_intronlist_from_annotation.py
  1. b - get_intron_features.py
  2. evaluate_features.py
  3. find_optimal_param_set.py

1 (a) filter_alignment.py

This tool filters a given SAM file according to specified filtering criteria. You can either pipe a SAM file into the script or provide an input file. Output can be written to stdout or a given file. By default the output file is modified version of the input name, tagged with the filtering criteria.

To specify filtering criteria you can choose between maximal mismatches, min exon length (which is the minimal segment length of spliced reads), a maximal intron length, or features such as the clipping state.

Further details are written in the usage screen of the script. Just type:

python filter_alignment.py

to print the usage screen.

1 (b) filter_features.py

The script filters a given feature list according to given criteria. Only by using this script you can later filter according to minimal splice site coverage with filter_alignment.py (see section 2.1(a)).

To create the feature list of an alignment, you can run get_intron_features.py (section 2.2(b)).

Further details are written in the usage screen of the script. Just type:

python filter_features.py

to print the usage screen.

2 (a) gen_intronlist_from_annotation.py

This tool generate the internal intron representation of annotated introns from any given annotation in GFF3 format. You only have to specify the path to the annotation file using the option -a and an output file via the option -o.

The script requires BioPython to be available in the PYTHONPATH. Please use section 1 of this file for further information.

Further details are written in the usage screen of the script. Just type:

python gen_intronlist_from_annotation.py

to print the usage screen.

2 (b) get_intron_features.py

With this script you can extract the intron positions together with descriptive alignment features into a so called feature-file.

You only need to specify an input alignment in SAM or BAM format using the option -a and the output destination using -o. If the input is BAM you must switch the input type using -b. For BAM input you further need SAMtools installed (see section 1). You either need SAMtools in your PATH or you have to provide the -s option.

Further details are written in the usage screen of the script. Just type:

python get_intron_features.py

to print the usage screen.

3 evaluate_features.py

This script evaluates the agreement of a given alignment to a given annotation that is represented as an intron list. First you need to generate the intron list (see section 2.2a) and the alignment feature file (section 2.2b).

With the option -i you can choose what annotation intron list to use and with -f you provide the alignment feature file.

Additionally, you can filter the evaluated features by many different criteria as maximal mismatches or intron length. You can also restrict the evaluation to several chromosomes/conttigs or to specific introns.

Further details are written in the usage screen of the script. Just type:

python evaluate_features.py

to print the usage screen.

4 find_optimal_param_set.py

This script performs an exhaustive search over the three parameters coordinates
  • max mismatches [0, 1, 2, 3, 4, 5, 6]
  • min segment length [2, 4, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18]
  • min intron read support [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]

The tested values for each criterion are currently hard coded and are written in brackets behind the criterion. If you want to change them manually, just search for the term "SEARCH SPACE" in find_optimal_param_set.py.

As for the feature evaluation: with the option -i you can choose what annotation intron list to use and with -f you provide the alignment feature file.

You further need to define two output files:

-b , --best_score   the file to store the best scoring parameters
-m , --matrix       the file to store all test parameter combinations

Further details are written in the usage screen of the script. Just type:

python find_optimal_param_set.py

to print the usage screen.

Document Actions