Test-drive the GATK tools and Best Practices pipelines on Terra
Check out this blog post to learn how you can get started with GATK and try out the pipelines in preconfigured workspaces (with a user-friendly interface!) without having to install anything.
Creating Amplicon Sequences - RETIRED
Please note that this article has not been updated in a very long time and may no longer be applicable. Use at your own risk.
Creating Amplicon Sequences
Note that earlier versions of the GATK used a different tool.
For a complete, detailed argument reference, refer to the GATK document page here
This tool generates amplicon sequences for use with the Sequenom primer design tool. The output of this tool is fasta-formatted, where the characters [A/B] specify the allele to be probed (see Validation Amplicons Output further below). It can mask nearby variation (either by 'N' or by lower-casing characters), and can try to restrict sequenom design to regions of the amplicon likely to generate a highly specific primer. This tool will also flag sites with properties that could shift the mass-spec peak from its expected value, such as indels in the amplicon sequence, SNPs within 4 bases of the variant attempting to be probed, or multiple variants selected for validation falling into the same amplicon.
Lowercase and Ns
Ns in the amplicon sequence instructs primer design software (such as Sequenom) not to use that base in the primer: any primer will fall entirely before, or entirely after, that base. Lower-case letters instruct the design software to try to avoid using the base (presumably by applying a penalty for doing so), but will not prevent it from doing so if a good primer (i.e. a primer with suitable melting temperature and low probability of hairpin formation) is found.
ValidationAmplicons relies on the GATK Sting BWA/C bindings to assess the specificity of potential primers. The wiki page for Sting BWA/C bindings contains required information about how to download the appropriate version of BWA, how to create a BWT reference, and how to set your classpath appropriately to run this tool. If you have not followed the directions to set up the BWA/C bindings, you will not be able to create validation amplicon sequences using the GATK. There is an argument (see below) to disable the use of BWA, and lower repeats within the amplicon only. Use of this argument is not recommended.
Running Validation Amplicons
Validation Amplicons requires three input files: a VCF of alleles you want to validate, a VCF of variants you want to mask, and a Table of intervals around the variants describing the size of the amplicons. For instance:
Alleles to Validate
##fileformat=VCFv4.0 #CHROM POS ID REF ALT QUAL FILTER INFO 20 207414 . G A 85.09 PASS . // SNP to validate 20 792122 . TCCC T 22.24 PASS . // DEL to validate 20 994145 . G GAAG 48.21 PASS . // INS to validate 20 1074230 . C T 2.29 QD . // SNP to validate (but filtered) 20 1084330 . AC GT 42.21 PASS . // MNP to validate
HEADERpos name 20:207334-207494 20_207414 20:792042-792202 20_792122 20:994065-994225 20_994145 20:1074150-1074310 20_1074230 20:1084250-1084410 20_1084330
Alleles to Mask
##fileformat=VCFv4.1 #CHROM POS ID REF ALT QUAL FILTER INFO 20 207414 . G A 77.12 PASS . 20 207416 . A AGGC 49422.34 PASS . 20 792076 . A G 2637.15 HaplotypeScore . 20 792080 . T G 161.83 PASS . 20 792087 . CGGT C 179.84 ReadPosRankSum . 20 792106 . C G 32.59 PASS . 20 792140 . C G 409.75 PASS . 20 1084319 . T A,C 22.24 PASS . 20 1084348 . TACCACCCCACACA T 482.84 PASS .
Validation Amplicons Output
The output from Validation Amplicons is a fasta-formatted file, with a small adaptation to represent the site being probed. Using the test files above, the output of the command
java -jar $GATK/dist/GenomeAnalysisTK.jar \ -T ValidationAmplicons \ -R /humgen/1kg/reference/human_g1k_v37.fasta \ -BTI ProbeIntervals \ --ProbeIntervals:table interval_table.table \ --ValidateAlleles:vcf sites_to_validate.vcf \ --MaskAlleles:vcf mask_sites.vcf \ --virtualPrimerSize 30 \ -o probes.fasta \ -l WARN
>20:207414 INSERTION=1,VARIANT_TOO_NEAR_PROBE=1, 20_207414 CCAACGTTAAGAAAGAGACATGCGACTGGGTgcggtggctcatgcctggaaccccagcactttgggaggccaaggtgggc[A/G*]gNNcacttgaggtcaggagtttgagaccagcctggccaacatggtgaaaccccgtctctactgaaaatacaaaagttagC >20:792122 Valid 20_792122 TTTTTTTTTagatggagtctcgctcttatcgcccaggcNggagtgggtggtgtgatcttggctNactgcaacttctgcct[-/CCC*]cccaggttcaagtgattNtcctgcctcagccacctgagtagctgggattacaggcatccgccaccatgcctggctaatTT >20:994145 Valid 20_994145 TCCATGGCCTCCCCCTGGCCCACGAAGTCCTCAGCCACCTCCTTCCTGGAGGGCTCAGCCAAAATCAGACTGAGGAAGAAG[AAG/-*]TGGTGGGCACCCACCTTCTGGCCTTCCTCAGCCCCTTATTCCTAGGACCAGTCCCCATCTAGGGGTCCTCACTGCCTCCC >20:1074230 SITE_IS_FILTERED=1, 20_1074230 ACCTGATTACCATCAATCAGAACTCATTTCTGTTCCTATCTTCCACCCACAATTGTAATGCCTTTTCCATTTTAACCAAG[T/C*]ACTTATTATAtactatggccataacttttgcagtttgaggtatgacagcaaaaTTAGCATACATTTCATTTTCCTTCTTC >20:1084330 DELETION=1, 20_1084330 CACGTTCGGcttgtgcagagcctcaaggtcatccagaggtgatAGTTTAGGGCCCTCTCAAGTCTTTCCNGTGCGCATGG[GT/AC*]CAGCCCTGGGCACCTGTNNNNNNNNNNNNNTGCTCATGGCCTTCTAGATTCCCAGGAAATGTCAGAGCTTTTCAAAGCCC
Note that SNPs have been masked with 'N's, filtered 'mask' variants do not appear, the insertion has been flanked by Ns, the unfiltered deletion has been replaced by Ns, and the filtered site in the validation VCF is not marked as valid. In addition, bases that fall inside at least one non-unique 30-mer (meaning no multiple MQ0 alignments using BWA) are lower-cased. The identifier for each sequence is the position of the allele to be probed, a 'validation status' (defined below), and a string representing the amplicon. Validation status values are:
Valid // amplicon is valid SITE_IS_FILTERED=1 // validation site is not marked 'PASS' or '.' in its filter field ("you are trying to validate a filtered variant") VARIANT_TOO_NEAR_PROBE=1 // there is a variant too near to the variant to be validated, potentially shifting the mass-spec peak MULTIPLE_PROBES=1, // multiple variants to be validated found inside the same amplicon DELETION=6,INSERTION=5, // 6 deletions and 5 insertions found inside the amplicon region (from the "mask" VCF), will be potentially difficult to validate DELETION=1, // deletion found inside the amplicon region, could shift mass-spec peak START_TOO_CLOSE, // variant is too close to the start of the amplicon region to give sequenom a good chance to find a suitable primer END_TOO_CLOSE, // variant is too close to the end of the amplicon region to give sequenom a good chance to find a suitable primer NO_VARIANTS_FOUND, // no variants found within the amplicon region INDEL_OVERLAPS_VALIDATION_SITE, // an insertion or deletion interferes directly with the site to be validated (i.e. insertion directly preceding or postceding, or a deletion that spans the site itself)
Warnings During Traversal
The files provided to Validation Amplicons should be such that all generated amplicons are valid. That means:
There are no variants within 4bp of the site to be validated There are no indels in the amplicon region Amplicon windows do not include other sites to be probed Amplicon windows are not too short, and the variant therein is not within 50bp of either edge All amplicon windows contain a variant to be validated Variants to be validated are unfiltered or pass filters
The tool will warn you each time any of these conditions are not met.