Skip to content

Latest commit

 

History

History

flankTE

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 
 
 

flankTE

Scripts to extract SNPs in the flanking regions of TE insertions. SNPs are extracted from whole-genome consensus sequences of the sample WGS datasets used for the TE calling analyses.

There are two types of the scripts.

  1. flankTE-SNP.py extract various types of SNP that support specific groupings of the samples, that need to be configured in the script itself.
  2. flankTE-tree.py Extract SNPs that according to various branches on a species tree. This also requires configuration in the script file.

Dependencies

Resolve dependencies i.e. using pip

pip install biopython pyfaidx dendropy ete2

flankTE runs on Python 2.7

Usage

python flankTE-tree.py 
usage: flankTE-tree.py [-h] -i FASTA -f FLANK [-o OVERLAPPING]
                                [-w WINDOWSIZE] -m MATRIX

Command line arguments

  • -i Whole-genome consensus sequence (fasta format), supplied for each taxa that is analysed, e.g. -i A.cons.fa -i B.cons.fa ...
  • -f Size of flanking regions in base pairs (i.e. 10000). Note that the first 500 bp surrounding the breakpoint will not be analysed in order to account for spuriously mapped reads.
  • -o True/[False]. Indicate whether the window size should be extended in each round (True) or if windows should be distinct regions of size [windowssize] (This is the default)
  • -w Window size in base pairs [1000]
  • -m Tab-separated files with presence information on TE insertions (see below for details).

Overlapping windows

The default behaviour is to count SNPs in distinct windows of size W surrounding the insertion site (breakpoint).

|<--W-4-->|<--W-3-->|<--W-2-->|<--W-1-->|<500bp omitted><BREAKPOINT><500bp omitted>|<--W+1-->|<--W+2-->|<--W+3-->|<--W+4-->|

With -o True the program returns the increasing window sizes. This is most meaningful if the window size equals the flanking size, i.e. only a specific window around the breakpoint is investigated.

Window 1:         |<--W-->|<500bp omitted><BREAKPOINT><500bp omitted>|<--W-->|

Window 2: |<------W------>|<500bp omitted><BREAKPOINT><500bp omitted>|<------W------>|

etc...

Format of matrix file

chr1    10000  10001    sampleA,sampleB
chr1    20000  20001    sampleA,sampleC
...