NGS QC metrics for variant calling include coverage depth (>30x), base quality scores (Q30>80%), mapping quality, strand bias, variant allele frequency, and contamination estimates, ensuring CLIA-compliant diagnostic accuracy and reproducibility.
Pre-Analytical QC Metrics
Sample Quality Assessment
- DNA concentration - >10 ng/μL recommended
- A260/280 ratio - 1.8-2.0 for purity
- Fragment size - DIN >7 or RIN >7
- Contamination check - No bacterial/fungal
- Sample identity - STR profiling/fingerprinting
Library Preparation QC
Library QC Thresholds:
- Insert size: 150-500bp (WES/WGS)
- Molarity: >2nM final
- Adapter contamination: <0.1%
- Library complexity: >70% unique
- Balanced indices: <2x variation
Sequencing Run Metrics
Illumina Quality Scores
| Metric | Threshold | Clinical Requirement |
|---|---|---|
| Cluster Density | 170-220 K/mm² | Optimal loading |
| Clusters PF | >80% | High quality |
| Q30 Bases | >80% | CLIA standard |
| Error Rate | <1% | Accuracy requirement |
| PhiX Alignment | 0.3-1% | Control spike-in |
Coverage Uniformity
- Mean coverage - >100x for somatic, >30x germline
- Coverage uniformity - >90% bases at 20x
- Target coverage - >95% regions covered
- Duplicate rate - <20% for WES, <5% WGS
- On-target rate - >70% for capture methods
Alignment Quality Metrics
Mapping Statistics
- Mapped reads - >95% properly paired
- MAPQ scores - >90% with MAPQ≥20
- Insert size distribution - Gaussian peak
- Chimeric reads - <5% of total
- Soft-clipped bases - <10% concern threshold
Contamination Assessment
VerifyBamID Output:
- FREEMIX <0.02 (2% contamination)
- CHIPMIX <0.02 for related samples
- Sample swaps via genotype concordance
Variant Calling QC
Variant Quality Metrics
- QUAL score - Phred-scaled confidence
- QD (Quality by Depth) - QUAL/DP >2
- FS (Fisher Strand) - <60 for SNPs, <200 INDELs
- MQ (Mapping Quality) - >40 average
- MQRankSum - >-12.5 mapping quality bias
Allele-Specific Metrics
| Metric | Description | Threshold |
|---|---|---|
| VAF | Variant allele frequency | >5% somatic, >20% germline |
| AD | Allelic depth | >5 alt reads |
| DP | Total depth | >30x clinical |
| SB | Strand bias | Fisher p>0.001 |
| AB | Allele balance | 0.25-0.75 for het |
CLIA Validation Parameters
Analytical Performance
- Sensitivity - Detection at stated VAF
- Specificity - False positive rate <1%
- Precision - Inter/intra-run CV <5%
- Accuracy - Concordance with orthogonal method
- LOD/LOQ - Limit of detection quantification
Reference Materials
Validation Samples:
- NA12878 (Genome in a Bottle)
- HD200/HD300 (Horizon Discovery)
- Seraseq (SeraCare)
- CAP proficiency samples
- Clinical validation cohort (>30)
Bioinformatics Pipeline QC
Pipeline Components
- FastQC - Read quality assessment
- MultiQC - Aggregate reporting
- Picard Tools - Alignment metrics
- GATK - Variant calling QC
- VCFtools - Variant statistics
Automated QC Checks
- Sample tracking - LIMS integration
- Contamination flags - Auto-detection
- Coverage alerts - Below threshold warnings
- Variant annotation - ClinVar, COSMIC
- Report generation - Clinical interpretation
Throughput Considerations
Batch Effects
- Batch design - Randomized layout
- Control samples - Per batch inclusion
- Normalization - Batch correction methods
- Trend monitoring - Control charts
- Root cause analysis - Failure investigation
Performance Monitoring
Key Performance Indicators:
├── TAT (turnaround time)
├── First-pass rate >95%
├── Repeat rate <5%
├── Critical value reporting
└── Physician satisfaction
Comprehensive QC metrics throughout the NGS workflow ensure diagnostic-grade variant calling accuracy, supporting clinical decision-making while meeting regulatory requirements for precision medicine applications.