WGS: Whole genome sequencing data analysis

Installation:

git clone https://github.com/YaoZhou89/WGSc.git
cd WGSc
make

Quality control of SNPs

The multiple polyploid levels and low coverage of depth may cause false positives in SNP calling due to misalignment and sequencing errors, so we applied a specific pipeline with six filters to eliminate low-confidence SNPs. The low-confidence SNPs were identified and removed by either technological or genetic filters. Technically, SNPs with low mapping quality, biased depth distribution or potential sequencing error were discarded. Genetically, SNPs that broke genetic laws, such as linkage disequilibrium or identity by descent (IBD), were discarded.

Quality filter

Usage:

java -jar WGS.jar --model vcf --type quality --MQ 30 --FS 60 --MQRankSum -12.5 --ReadPosRankSum --BSQRankSum 0 --SOR 3 --file in.vcf --out out.vcf

Parameters:

MQ: MQ tag in vcf file, if the value less than the threshold, the SNP will be eliminated.
FS： FS tag in vcf file, if the value larger than the threshold, the SNP will be eliminated.
MQRankSum: MQRankSum in vcf file, if the value less than the threshold, the SNP will be eliminated.
ReadPosRankSum: ReadPosRankSum tag,in vcf file, if the value less than the threshold, the SNP will be eliminated.
BSQRankSum: BSQRankSum tag in vcf file, if the value less than the threshold, the SNP will be eliminated.
SOR: SOR tag in vcf file, if the value larger than the threshold, the SNP will be eliminated.

Depth filter

Usage:

java -jar WGS.jar --model vcf --type DepthFilter --maxSD 3 --min-depth 10 --max-depth 50 --file in.vcf --out out.vcf

Parameters:

maxSD: standard deviation(SD) of the depth for all individuals, if the SD larger than the threshold, the SNP will be eliminated.
min-depth: minimum depth allowed for the total depth.
max-depth: maximum depth allowed for the total depth.

Tips: Manually determine the appropriate value for these parameters based on the distribution of total depth.

Segregation test filter

Usage:

java -jar WGS.jar --model vcf --type ST --file in.vcf.gz --out out.vcf

Tips： the input file should be compressed format and also indexed with tabix.

Linkage disequilibrium (LD) filter

Usage:

java -jar WGS.jar --model vcf --type LDfilter --windowSize 50 --threshold 0.01 --file in.vcf.gz --out out.vcf

Parameters:

windowSize: number of SNPs consider;
threshold: SNPs with mean correlation less than the threshold will be eliminated.

Tips： the input file should be compressed format and also indexed with tabix.

IBD filter

Usage:

java -jar WGS.jar --model vcf --type IBDfilter --anchorFile anchorSNPs.vcf --minComp 200 --maxIBDDist 0.03 --windowSize 2000 --numThreads 32 --file in.vcf --out outprefix

Parameters

anchorFile: file name of anchor SNPs, in vcf format
minComp: minimum number of comparable anchor sites
maxIBDDist: If the genetic distance within such a window was not larger than the given value, the 2 lines were considered to be in IBD. 
windowSize: the number of SNPs in a window 
numThreads: number of threads

Tips: Manually determine the maxIBDDist value based on the the mean distance across all pairs (suggested about 10 times smaller than the mean value using in the maize hapmap3)

Minor allele count filter

Usage: using vcftools

vcftools --vcf in.vcf --recode --mac 3 --out outprefix

Author

Dr. Yao Zhou (zhouyao@caas.cn)