phil9s / variant_filter

Series of scripts for filtering variant calling format files

Variant Filtering Script - R & Bash

A set of scripts for filtering multi-sample VCF files to generate candidate variant lists

Main command:

./variant_filtering.sh --project my_project \
--input /data/Alignments/your_merged_vcf.vcf \
--output /home/user/ \
--config variant_filtering.config \
--temp TRUE

1. variant_filtering_results.tsv

Ouput description table

Column Name	Description
Id	Variant Id Number
CHROM	Chromosome number
POS	Variant loci position
rsID	rsID or position again if not known
REF	Reference base
ALT	Alternate base
QUAL	QUAL score
GENE	Gene symbol
TYPE	Exonic/splicing - what type of region
AA	AA change for multiple transcripts - look up appropriate one
CONSEQUENCE	Functional change or consequence (stop gain etc)
X1000G	Rarity in 1000 genomes
EXAC	Rarity in ExAC consortium
CADD	CADD score variant damage prediction
HET_val	Number of samples called with a het alt call at this position
HOM_val	Number of samples called with a hom alt call at this position
MISS_rate	The pecentage rate of sites failing to genotype (missing rate)
AC_all	Alt allelic counts for this site (post filtering) - homozygous = 2 towards count
AC_nsyn	Same as AC_all but only for nonsynonymous changes
AC_trnc	Same as AC_all but only for truncating changes (stop gain, frameshifts)
AC_other	Same as AC_all but only for all other types of changes
AggAF_Trunc	Aggregated allele frequency of truncating mutations based on 1000Genomes or ExAc
AggAF_nsyn	Same as AggAF_Trunc but for nonsynonymous mutations
RVIS_Pct	Residual variant intolerance score percentile - Higher is more tolerant
GDIS_Phred	Genetic damage intolerance score - Phred scaled - Higher is more tolerant
Sample_Id	Columns from this point are genotyping columns - 0 = ref, 1 = het, 2 = hom, -9 = missing

2. variant_filtering_GeneAC.tsv

• Seperate file containing only the AC_* columns from variant_filtering_results.tsv
• Used for statistical analysis of case/controls for all genes present in the variant file

3. variant_filtering_results_AD.tsv

• Allelic depth ratio matrix
• describes the proportion of reads supporting the REF against ALT
• Calculates as ALT READs / REF reads + ALT reads
• Used as a filtering step in variant_filtering.config file - variant retained if one sample passes the threshold
• Multi-allelic variants are omitted from this filter for ease of parsing the data

4. R_log.txt

• Log file of each filtering step in the R filtering script
• Denotes start times and variant counts at each stage
• Declares the number of multi-allelic variants carried forward
• Describes the number of variants in which AC_all was not equal to sum(AC_trunc + AC_nsyn + AC_other) Should be a very low proportion of total variants and is due to unusual variant type/consequence combinations

5. variantfilter.log

• Log file for the bash portion of the variant filtering script
• Contains the parameters used and declares which thresholds were specified in the variant_filtering.config file
• Mainly just for error checks and confirming the script has completed

6. variant_comphet-biallelic_results.tsv

• Calculates sets of variants that are potentially compound heterozygous
• Sites are retained if two or more occur in the same sample and gene simultanously

Required dependencies and software:

• Annovar - Including various databases
• vcftools
• bcftools
• R (version 3.4.0)

Required R Libraries:

• stringr