Pre-SCNAClonal

Introduction

Tumor samples often present heterogeneity, containing not only multiple subpopulations of cancerous cells defined by distinct somatic mutations, but also the normal, non-cancerous cells. While using the whole-genome sequencing data of tumor sample to reconstruct the subclonal composition determined by somatic copy number alternations (SCNAs), the absolute copy number and its subclonal population are both required to be estimated. However, the raw reads in WGS data present coupling GC bias in the same SCNA genome segmentation of paired tumor and normal samples, which largely affect the absolute copy number estimation conducted by existing SCNV based subclonal inferring tools. We provide Pre-SCNVClonal, a comprehensive package for both automatically and visually correcting coupling GC bias and baseline correction. Pre-SCNVClonal could be strung together with the SCNV based subclonal inferring tool as a pipeline or run individually as needed.

Requirement

• Python2.7

• python package:

• scipy

• numpy

• pickle

• matplotlib

Manual for Visually correct GC bias

Manually adjust the linear regression line

1. Click this button to pan axes with left mouse, zoom with right mouse

2. Adjust the alpha and area bars to better visualize the stripes

3. Adjust the Slope delta and Interception delta bars to move linear regression line overlope one of the stripes

4. Click OK button

Manually eliminate the horizontal GC bias

1. Click this button off after the adjustment of alpha and area bars

2. Hold left mouse and move to pick out one stripe out, then Pre-SCNAClonal with redraw the linear regression line.

3. Click OK button

Usage of Pre-SCNAClonal

Generally, SCNAs based tumor composition reconstruction tools utilize the read count, B allele frequency (BAF) information to estimate the abslute copy number and tumor subclonal frequency. According to the method of geting read count and BAF information, these reconstruction tools could be grouped into two classes, 1) Get the read count and BAF information totally from the BAM files, i.e. MixClone; 2) Get the read count and BAF position from heterozygous SNP position, then get the BAF information from BAM files at the heterozygous SNP position. i.e. THetA;

Example of MixClone like

usage

usage: run.py MixClone [-h] [--pkl_path PKL_PATH]
		       [--max_copynumber MAX_COPYNUMBER]
		       [--subclone_num SUBCLONE_NUM]
		       [--baseline_thred_LOH BASELINE_THRED_LOH]
		       [--baseline_thred_APM BASELINE_THRED_APM]
		       [--pkl_flag PKL_FLAG] [--min_depth MIN_DEPTH]
		       [--min_base_qual MIN_BASE_QUAL]
		       [--min_map_qual MIN_MAP_QUAL]
		       [--process_num PROCESS_NUM]
		       [--gc_correction_method GC_CORRECTION_METHOD]
		       normal_bam tumor_bam reference_genome
		       input_filename_base segments_bed BICseq_bed_corrected

positional arguments:
  normal_bam            BAM file for normal sample.
  tumor_bam             BAM file for tumor sample.
  reference_genome      FASTA file for reference genome.
  input_filename_base   Base name of the preprocessed input file to be
			created.
  segments_bed          BED file for segments.
  BICseq_bed_corrected  The name of corrected BICseq result file

optional arguments:
  -h, --help            show this help message and exit
  --pkl_path PKL_PATH   Load the pkl path
  --max_copynumber MAX_COPYNUMBER
			Set the maximum copy number
  --subclone_num SUBCLONE_NUM
			Set the subclone number
  --baseline_thred_LOH BASELINE_THRED_LOH
			The threshold of LOH sites fraction within each
			segment to define the segment is LOH, the range is
			[baseline_thred_LOH, 1]. Default is 0.16.
  --baseline_thred_APM BASELINE_THRED_APM
			The threshold of average P and M SNP sites fraction
			within each segment to define the segment as baseline,
			the range is [baseline_thred_APM, 1]. Default is 0.2.
  --pkl_flag PKL_FLAG   The pkl flag
  --min_depth MIN_DEPTH
			Minimum reads depth required for both normal and tumor
			samples. Default is 20.
  --min_base_qual MIN_BASE_QUAL
			Minimum base quality required. Default is 10.
  --min_map_qual MIN_MAP_QUAL
			Minimum mapping quality required. Default is 10.
  --process_num PROCESS_NUM
			Number of processes to launch for preprocessing.
			Default is 1.
  --gc_correction_method GC_CORRECTION_METHOD
			The gc correction method, one of auto and visual

####Example

python2.7 $PRESCNVCLONAL_PATH/run.py MixClone \
    $normal_bam \
    $tumor_bam \
    $reference_genome \
    $input_filename_base \
    $segments_bed \
    $BICseq_bed_corrected\
    --pkl_path $pkl_path\
    --max_copynumber $max_copynumber \
    --subclone_num $subclone_num \
    --baseline_thred_LOH $baseline_thred_LOH \
    --baseline_thred_APM $baseline_thred_APM \
    --min_depth $min_depth  \
    --min_base_qual $min_base_qual \
    --min_map_qual $min_map_qual \
    --process_num $process_num \
    --gc_correction_method auto \

Example of THetA like

Usage

usage: run.py THetA [-h] [--pkl_path PKL_PATH] [--pkl_flag PKL_FLAG]
		    [--max_copynumber MAX_COPYNUMBER]
		    [--subclone_num SUBCLONE_NUM]
		    [--sampleNumber SAMPLENUMBER]
		    [--gc_correction_method GC_CORRECTION_METHOD]
		    [--baseline_thred_LOH BASELINE_THRED_LOH]
		    [--baseline_thred_APM BASELINE_THRED_APM]
		    [--baseline_thred_SNPDENSITY BASELINE_THRED_SNPDENSITY]
		    [--gamma GAMMA] [--process_num PROCESS_NUM]
		    input_filename_base BICseq_bed BICseq_bed_corrected
		    tumor_SNP normal_SNP

positional arguments:
  input_filename_base   Base name of the preprocessed input file to be
			created.
  BICseq_bed            BICseq result file
  BICseq_bed_corrected  The name of corrected BICseq result file
  tumor_SNP             tumor snp file, generated by THetA script
  normal_SNP            normal snp file. generated by THetA script

optional arguments:
  -h, --help            show this help message and exit
  --pkl_path PKL_PATH   Load the pkl path
  --pkl_flag PKL_FLAG   The pkl flag
  --max_copynumber MAX_COPYNUMBER
			Set the maximum copy number
  --subclone_num SUBCLONE_NUM
			Set the subclone number
  --sampleNumber SAMPLENUMBER
			Set the sample number for visual plot
  --gc_correction_method GC_CORRECTION_METHOD
			The gc correction method, one of auto and visual
  --baseline_thred_LOH BASELINE_THRED_LOH
			The threshold of LOH sites fraction within each
			segment to define the segment is LOH, the range is
			[baseline_thred_LOH, 1]. Default is 0.3.
  --baseline_thred_APM BASELINE_THRED_APM
			The threshold of average P and M SNP sites fraction
			within each segment to define the segment as baseline,
			The range is [baseline_thred_APM, 1]. Default is 0.01.
  --baseline_thred_SNPDENSITY BASELINE_THRED_SNPDENSITY
			The threshold of density of SNP sites in segment, The
			range is [baseline_thred_APM, 1]. Default is 0.01.
  --gamma GAMMA         gamma parameter for determine heterozygous snp loci
  --process_num PROCESS_NUM
			Number of processes to launch for preprocessing.
			Default is 1.

Example

python2.7 ./run.py THetA \
    $segments_bed \
    $BICseq_bed_corrected\
    $tumor_SNP_counts \
    $nomal_SNP_counts \
    --gc_correction_method [visual|auto]\
    --baseline_thred_LOH 0.16 \
    --baseline_thred_APM 0.6