top of page

Whole Genome In-Depth Variant Calling

Updated: Sep 19, 2023

Exploring Short Variants, Structural Variations, and Copy Number Variations in Human Genomes.

This study offers a comprehensive analysis of whole genome variant calling, employing the rigor of GATK best practices to scrutinize 100 human samples. Encompassing short variants (SNVs), structural variations (SVs), and copy number variations (CNVs), the research unveils a nuanced perspective on the complexity of human genomics, enhancing our understanding of underlying genetic patterns.

Each step within the pipeline is designed to ensure the accuracy and reliability of the findings.

By executing quality control, duplication marking, and recalibration of base qualities and variant quality scores, we establish a robust foundation for the analysis. This stringent approach minimizes the risk of false positive variants, a critical aspect in genomic research where precision is of utmost significance. Fusing these steps safeguards against inaccuracies, enabling us to extract meaningful genetic insights confidently.

Technical Analysis Details:

  • Quality Control of Raw Reads: Initiating the analysis, i perform quality control on raw sequencing data, ensuring high-quality reads for accurate downstream analysis.

  • Mapping to Reference Genome: The processed reads are mapped to a reference genome, establishing the genomic coordinates for each read.

  • Marking Duplicated Reads: Identifying and marking duplicated reads is crucial to ensure accurate variant calling, as duplicates can skew variant counts.

  • Recalibrating Base Quality Scores: We employ base quality score recalibration, refining the accuracy of variant calls by compensating for systematic errors in base quality predictions.

  • Calling Variants in GVCF Mode: Leveraging the power of the HaplotypeCaller tool in GVCF mode, we call variants per sample to generate intermediate files in GVCF format, facilitating efficient joint genotyping.

  • Consolidate GVCFs: We utilize GenomicsDBImport to consolidate GVCF files from multiple samples, enhancing scalability and streamlining the subsequent joint genotyping step.

  • Filter Variants and Refine Genotypes: Through GenotypeGVCFs, we perform joint genotyping, enabling cohort-wide analysis and producing a squared-off matrix of genotypes across samples. This step empowers sensitive variant detection and produces genotype information crucial for downstream analyses.

  • Annotate Variants: Our analysis includes comprehensive variant annotation, which adds biological context to each variant, enriching the understanding of potential functional impact.

The GATK Variant Quality Score Recalibration (VQSR) model is a robust statistical framework meticulously designed to refine the accuracy of variant calls.

Using machine learning techniques, VQSR identifies annotation profiles associated with high-confidence variants, generating a more reliable quality score known as VQSLOD. This quality score outperforms the traditional QUAL score by offering a calibrated probability estimate for each variant's validity. The Gaussian distribution plots produced during this process visually represent the model's calibration performance. These plots and the Tranches plot serve as invaluable tools in interpreting the model's effectiveness, guiding researchers in selecting quality thresholds that balance specificity and sensitivity. In essence, the GATK VQSR model enhances the precision of variant calling, ensuring that the reported genetic variations are of the highest quality and minimizing the risk of false positive calls.

The "VariantRecalibrator" step involves fitting a Gaussian mixture model to contextual annotations of variants. This model is trained using variants known as true-positives, allowing the assignment of probabilities to new variants. A report is generated for visualization, depicting the fitting of the probability model to data by projecting the Gaussian mixture model in 2D for each annotation combination used in modeling.

The Transition/Transversion (Ti/Tv) plot examines the Transition to Transversion SNPs ratio, specifically in the context of whole genome sequencing data utilized in this project. In a random scenario, this ratio would be approximately 0.5 due to the prevalence of potential transversions. However, this ratio can rise to around 2.01 due to biological processes like cytosine deamination. Detecting significant deviations from these expected values could indicate artifactual variants, potentially signaling an excess of false positives in the call set derived from the whole genome sequencing data.

A bar plot depicting Structural Variations (SVs) and Copy Number Variations (CNV) across distinct genomic regions, each section reflects the relative frequency within its corresponding genomic category, shedding light on the genomic landscape's dynamics and highlighting potential regions of interest.

You quickly deal with a mountain of variations when diving into variant calling analysis results. But here's where things get interesting – I've sifted through all this data and pulled out the significant bits. This section is all about sharing what I've found. You'll get your hands on valuable files, simple-to-follow plots, and reports that make sense.

Output and Results:


  • Comprehensive Variant Calling Format (VCF) file containing all identified variations.

  • BAM files representing aligned sequence data.

  • Annotated VCF file enriched with relevant information.

  • Filtered VCF file, refined through variant recalibration.

  • Detailed documentation of the GATK VQSR model.


  • Visual representation of variant distribution.

  • Potential Principal Component Analysis (PCA) plot generated from VCF data.

  • Gaussian plots are showcasing each filter's impact in the VQSR model.


  • Succinct presentation slides covering the entire analysis process with graphical insights.

  • A Comprehensive technical report outlining analysis tools and versions employed.


Go back to the 'Services' Page | Explore other projects

30 views0 comments

Recent Posts

See All

Genome Assembly

Introduction: Viral genome assembly from sequenced reads and identification of variants along the genome. Whole-genome sequencing (WGS) of pathogens is gaining vital significance, encompassing basic r

Single-cell RNA sequencing analysis

Single-cell RNA sequencing (scRNA-seq) is a revolutionary technique that enables researchers to explore the intricacies of cell types, states, and interactions across various human tissues. By providi

DNA methylation sequencing analysis

Whole Genome Methylation Sequencing Analysis. WGBS is a preeminent technique for exploring the methylome - the comprehensive methylation status of an entire genome. This method holds immense significa


bottom of page