Illumina FFPE DNA Prep with Exome 2.5
  • Illumina FFPE DNA Prep with Exome 2.5 Enrichment
  • Demo Data
  • Resource Files
  • Software Guides
    • DRAGEN v4.3.17 Recipe
    • BaseSpace Sequence Hub (BSSH)
    • Illumina Connected Analytics (ICA)
  • Additional Information
    • Homologous Recombination Deficiency (HRD)
    • Spike-in panels
    • Germline variants
    • Troubleshooting the analysis
Powered by GitBook
On this page
Export as PDF
  1. Additional Information

Troubleshooting the analysis

Potential changes to the analysis settings to improve performance when analyzing Illumina FFPE DNA Prep with Exome 2.5 Enrichment.

Why are the Tumor Mutation Burden (TMB) values not what I expected?

Low TMB values

Variants that do not meet the minimum depth (DP) threshold are excluded from the TMB calculation. The --vc-callability-tumor-thresh command line option specifies the threshold value. The default value is 50, and this value assumes at least 100x coverage. If raw coverage is lower than 100x and and tumor purity >= 80%, the threshold value can be decreased to 30, by specifying --vc-callability-tumor-thresh 30 as an additional argument.

High TMB values

DRAGEN outputs both TMB and NonSyn TMB metrics, where TMB includes all Filtered Variants in the calculation and NonSyn TMB includes Filtered Nonsynonymous Variants. One metric may be more appropriate than the other for a given comparison. For example, Seracare TMB reference samples include only nonsynonymous variants in their TMB metric.

Why do I have too many/too few PASSing somatic variants?

Somatic variants with a quality score less than the threshold value [0..30] are marked as filtered in the output vcf. The --vc-sq-filter-thresholdcommand line option specifies the threshold value. The default value for Tumor/Normal mode is 17.5. We observed that with this library preparation kit, the threshold needed to be lowered slightly to 15 to achieve high sensitivity. Raise this value to improve specificity at the cost of sensitivity, or lower it to improve sensitivity at the cost of specificity.

Why do some qc-coverage metrics report different values between command line and BSSH?

There are some rare instances where DRAGEN and BSSH apps have different default settings that might lead to slight difference in coverage calculations. For instance, the BSSH Enrichment App outputs a Aggregate Summary Metrics file that include all reads with MAPQ>=0 in the coverage calculations, while in DRAGEN command-line, MAPQ=0 reads are filtered. To include MAPQ=0 reads in DRAGEN command-line coverage, the flag --qc-coverage-filters-1 “mapq<0” can be added to the preprocessing recipe.

Why is the CNV VCF file empty?

The first thing to check is the coverage of the normal sample. Normal sample counts are added to PON, and if the counts from the normal sample are nearly empty, the GC correction stage will fail. In that case, DRAGEN will output an empty VCF.

PreviousGermline variants

Last updated 3 months ago