VarDictJava is a variant discovery program written in Java and Perl. It is a partial Java port of VarDict variant caller.
The original Perl VarDict is a sensitive variant caller for both single and paired sample variant calling from BAM files. VarDict implements several novel features such as amplicon bias aware variant calling from targeted
sequencing experiments, rescue of long indels by realigning bwa soft clipped reads and better scalability
than Java based variant callers.
Please cite VarDict:
Lai Z, Markovets A, Ahdesmaki M, Chapman B, Hofmann O, McEwen R, Johnson J, Dougherty B, Barrett JC, and Dry JR. VarDict: a novel and versatile variant caller for next-generation sequencing in cancer research. Nucleic Acids Res. 2016, pii: gkw227.
The link to is article can be accessed through: http://nar.oxfordjournals.org/cgi/content/full/gkw227?ijkey=Tk8eKQcYwNlQRNU&keytype=ref
Original coded by Zhongwu Lai 2014.
VarDictJava can run in single sample (see Single sample mode section), paired sample (see Paired variant calling section), or amplicon bias aware modes. As input, VarDictJava takes reference genomes in FASTA format, aligned reads in BAM format, and target regions in BED format.
- JDK 1.7 or later
- R language (uses /usr/bin/env R)
- Perl (uses /usr/bin/env perl)
- Internet connection to download dependencies using gradle.
###Getting source code
The VarDictJava source code is located at https://github.com/AstraZeneca-NGS/VarDictJava.
To load the project, execute the following command:
git clone --recursive https://github.com/AstraZeneca-NGS/VarDictJava.git
Note that original VardDict project is placed in this repository as submodule and it's contents can be found in sub-directory VarDict in VarDictJava working folder. So when you use
var2vcf_valid.pl. (see details and examples below), you have to add prefix VarDict:
The project uses Gradle and already includes a gradlew script.
To build the project, in the root folder of the project, run the following command:
./gradlew clean installApp
To generate Javadoc, in the build/docs/javadoc folder, run the following command:
./gradlew clean javadoc
###Single sample mode
To run VarDictJava in single sample mode, use a BAM file specified without the
| symbol and perform Steps 3 and 4 (see the Program workflow section) using
The following is an example command to run in single sample mode:
AF_THR="0.01" # minimum allele frequency <path_to_vardict_folder>/build/install/VarDict/bin/VarDict -G /path/to/hg19.fa -f $AF_THR -N sample_name -b /path/to/my.bam -z -c 1 -S 2 -E 3 -g 4 /path/to/my.bed | VarDict/teststrandbias.R | VarDict/var2vcf_valid.pl -N sample_name -E -f $AF_THR
VarDictJava can also be invoked without a BED file if the region is specified in the command line with
The following is an example command to run VarDictJava for a region (chromosome 7, position from 55270300 to 55270348, EGFR gene) with
<path_to_vardict_folder>/build/install/VarDict/bin/VarDict -G /path/to/hg19.fa -f 0.001 -N sample_name -b /path/to/sample.bam -z -R chr7:55270300-55270348:EGFR | VarDict/teststrandbias.R | VarDict/var2vcf_valid.pl -N sample_name -E -f 0.001 >vars.vcf
In single sample mode, output columns contain a description and statistical info for variants in the single sample. See section Output Columns for list of columns in the output.
###Paired variant calling
To run paired variant calling, use BAM files specified as
BAM1|BAM2 and perform Steps 3 and 4 (see the Program Workflow section) using
In this mode, the number of statistics columns in the output is doubled: one set of columns is for the first sample, the other - for second sample.
The following is an example command to run in paired mode:
AF_THR="0.01" # minimum allele frequency <path_to_vardict_folder>/build/install/VarDict/bin/VarDict -G /path/to/hg19.fa -f $AF_THR -N tumor_sample_name -b "/path/to/tumor.bam|/path/to/normal.bam" -z -F -c 1 -S 2 -E 3 -g 4 /path/to/my.bed | VarDict/testsomatic.R | VarDict/var2vcf_somatic.pl -N "tumor_sample_name|normal_sample_name" -f $AF_THR
The VarDictJava program follows the workflow:
- Get regions of interest from a BED file or the command line.
For each segment:
a. Find all variants for this segment in mapped reads:
i. Optionally skip duplicated reads, low mapping-quality reads, and reads having a large number of mismatches.
ii. Skip a read if it does not overlap with the segment.
iii. Preprocess the CIGAR string for each read.
iv. For each position, create a variant. If a variant is already present, adjust its count using the adjCnt function.
b. Realign some of the variants using special ad-hoc approaches.
c. Calculate statistics for the variant, filter out some bad ones, if any.
d. Assign a type to each variant.
e. Output variants in an intermediate internal format (tabular). Columns of the table are described in the Output Columns section.
**Note**: To perform Steps 1 and 2, use the Java program VarDict.
Perform a statistical test for strand bias using an R script.
Note: Use R script for this step.
- Transform the intermediate tabular format to VCF. Output the variants with filtering and statistical data.
Note: Use the Perl scripts
var2vcf_somatic.plfor this step.
Print help page
Print a header row decribing columns
Output splicing read counts
Do pileup regarless the frequency
Indicate the chromosome names are just numbers, such as 1, 2, not chr1, chr2
Debug mode. Will print some error messages and append full genotype at the end.
Indicate to remove duplicated reads. Only one pair with identical start positions will be kept
Indicate to move indels to 3-prime if alternative alignment can be achieved.
The hexical to filter reads. Default:
0x500(filter 2nd alignments and duplicates). Use
-F 0to turn it off.
Indicate whether the BED file contains zero-based cooridates, the same way as the Genome browser IGV does. -z 1 indicates that coordinates in a BED file start from 0. -z 0 indicates that the coordinates start from 1. Default:
1for a BED file or amplicon BED file. Use
0to turn it off. When using
-Roption, it is set to
Indicate it is amplicon based calling. Reads that do not map to the amplicon will be skipped. A read pair is considered to belong to the amplicon if the edges are less than int bp to the amplicon, and overlap fraction is at least float. Default:
Indicate whether to perform local realignment. Default:
1or yes. Set to
0to disable it.
-G Genome fasta
The reference fasta. Should be indexed (.fai). Defaults to:
The region of interest. In the format of chr:start-end. If chr is not start-end but start (end is omitted), then it is a single position. No BED is needed.
The delimiter for splitting
region_info, defaults to tab
The regular expression to extract sample names from bam filenames. Defaults to:
The sample name to be used directly. Will overwrite
The indexed BAM file
The column for chromosome
The column for the region start, e.g. gene start
The column for the region end, e.g. gene end
The column for a segment starts in the region, e.g. exon starts
The column for a segment ends in the region, e.g. exon ends
The column for a gene name, or segment annotation
The number of nucleotides to extend for each segment, default:
The threshold for allele frequency, default:
-r minimum reads
The minimum # of variance reads, default:
The minimum # of reads to determine strand bias, default:
If set, reads with mapping quality less than INT will be filtered and ignored
The phred score for a base to be considered a good call. Default: 25 (for Illumina). For PGM, set it to ~15, as PGM tends to underestimate base quality.
If set, reads with mismatches more than
INTwill be filtered and ignored. Gaps are not counted as mismatches. Valid only for bowtie2/TopHat or BWA aln followed by sampe. BWA mem is calculated as NM - Indels. Default: 8, or reads with more than 8 mismatches will not be used.
Trim bases after
[INT]bases in the reads
Extension of bp to look for mismatches after insersion or deletion. Default to 3 bp, or only calls when they're within 3 bp.
The read position filter. If the mean variants position is less that specified, it is considered false positive. Default: 5
For downsampling fraction, e.g.
70%downsampling. Default: No downsampling. Use with caution. The downsampling will be random and non-reproducible.
(good_quality_reads)/(bad_quality_reads+0.5). The quality is defined by
The reads should have at least mean
MapQto be considered a valid variant. Default: no filtering
The lowest frequency in a normal sample allowed for a putative somatic mutations. Defaults to
The indel size. Default: 120bp
The minimum matches for a read to be considered. If, after soft-clipping, the matched bp is less than INT, then the
read is discarded. It's meant for PCR based targeted sequencing where there's no insert and the matching is only the primers.
Default: 0, or no filtering
If this parameter is missing, then the mode is one-thread. If you add the -th parameter, the number of threads equals to the number of processor cores. The parameter -th threads sets the number of threads explicitly.
-VS STRICT | LENIENT | SILENT
How strict to be when reading a SAM or BAM.
STRICT- throw an exception if something looks wrong.
LENIENT- Emit warnings but keep going if possible.
LENIENT, only don't emit warning messages.
- Sample - sample name
- Gene - gene name from a BED file
- Chr - chromosome name
- Start - start position of the variation
- End - end position of the variation
- Ref - reference sequence
- Alt - variant sequence
- Depth - total coverage
- AltDepth - variant coverage
- RefFwdReads - reference forward strand coverage
- RefRevReads - reference reverse strand coverage
- AltFwdReads - variant forward strand coverage
- AltRevReads - variant reverse strand coverage
- Genotype - genotype description string
- AF - allele frequency
- Bias - strand bias flag
- PMean - mean position in read
- PStd - flag for read position standard deviation
- QMean - mean base quality
- QStd - flag for base quality standard deviation
- QRATIO - ratio of high quality reads to low-quality reads
- HIFREQ - variant frequency for high-quality reads
- EXTRAFR - Adjusted AF for indels due to local realignment
- SHIFT3 - No. of bases to be shifted to 3 prime for deletions due to alternative alignment
- MSI - MicroSattelite. > 1 indicates MSI
- MSINT - MicroSattelite unit length in bp
- NM - average number of mismatches for reads containing the variant
- HICNT - number of high-quality reads with the variant
- HICOV - position coverage by high quality reads
- 5pFlankSeq - neighboring reference sequence to 5' end
- 3pFlankSeq - neighboring reference sequence to 3' end
- SEGMENT:CHR_START_END - position description
- VARTYPE - variant type
The code is freely available under the MIT license.