This is the readme file for experimenting with a toy example using DECoNT.

This toy example is based on the real WES CNV calls of the sample HG00733 from 1000 Genomes Project.

1) Directory contains raw output file of XHMM (file with .xcnv extension).
2) Directory contains read depth counts of the sample used in the analysis. WES read counts obtained using sambamba v0.7.0. File is located in read_depths directory.
3) Directory contains an image of the expected output file, and the expected output file itself. Expected output file is named Expected_output_file.txt
4) Directory contains the main folder of DECoNT tool.


Steps:

0) Activate DECoNT environment using conda or get specified versions of the packages. Activation of DECoNT environment is explained in the main README file located in DECoNT\ Result\ Reproduction/ directory.

1) cd into /DECoNT/scripts/. There you will see DECoNT_polish.py script. This is the main polishing script of the tool DECoNT. The required arguments for this script are: --model, --callername, --input, --output, --samples. The optional arguments for this script are --gpu, --version, --help. For further detail regarding the arguments please refer to: https://github.com/ciceklab/DECoNT. 

2) DECoNT takes WES CNV calls produced by XHMM, CoNIFER, CODEX2 or Control-FREEC. In this tutorial we will use XHMM calls as input to DECoNT. It also requires read depth counts of the samples used in WES CNV calling. In this toy example, the raw output file of XHMM is given as HG00733_WES_CNV_CALLS_XHMM.xcnv and the read depth counts to the samples used in this analysis are given with the directory read_depths.


3) Run the command

$ python DECoNT_polish.py -m pretrained -cn XHMM -i ../../HG00733_WES_CNV_CALLS_XHMM.xcnv -o ../../ -s ../../read_depths/


With the above command we specify model weights as pertained (-m pretrained), caller name as XHMM (-cn XHMM), raw xhmm cnv call file as ../../HG00733_WES_CNV_CALLS_XHMM.xcnv (-i), directory to write the output file as ../../ and the read depth directory that contains read_depths regarding the samples in the analysis as ../../read_depths/ (-s).



After the script runs and completes the polishing process, the script will write the outputs of the polishing process to the directory specified with -o argument (DECoNT_XHMM_polished_cnvs.txt). You should expect the output file to be same as the file Expected_output_file.txt. 

Running time for this tutorial:
This toy example on an ordinary 2015 MacBook Pro computer with 8GB 1867 MHz DDR3 RAM and 2.7GHz Dual-Core Intel Core i5 processor, takes about ~9 minutes.
