Application Note

Benchmarking End-to-End, On-variants and results - GWAS

Premise, WGS Analysis, using Sentieon DNAscope Secondary Analysis with OmniTier Insight Tertiary Analysis

Introduction

Highlights

As sequencing technologies become faster and more affordable, a shift toward whole-exome analysis pipeline provides accurate and efficient WGS sequencing (WGS) is emerging. These approaches offer broader insights, enabling the discovery of novel genetic variants and a deeper understanding of complex traits, transforming research and personalized medicine in the genomic era.

Bioinformatics tools are also evolving to meet the requirements of challenging clinical applications. Initially, it was necessary to analyze whole genome data in the cloud, to take advantage of the massive compute resources available. However, as on-premise compute processing and AI technology has advanced, whole genome analysis is now available.

This application note details the benchmarking results of a whole genome data processing software pipeline comprising of Sentieon DNAscope, secondary analysis, followed by OmniTier Insight Hereditary, tertiary analysis, all running on OmniTier’s CompStor Workgroup, on-premise server. These software tools are designed to meet the high-performance requirements of whole genomes, as well as whole exomes and panels.

About Sentieon DNAscope

DNAscope uniquely integrates haplotype-based variant calling with machine learning to enhance accuracy. As the successor to GATK HaplotypeCaller6, DNAscope retains a similar logical framework while introducing key improvements in active region detection and local assembly, significantly boosting sensitivity and robustness, particularly in high-complexity regions.

By leveraging a machine learning model, DNAscope generates candidate variants with additional informative annotations. These annotated candidates are then processed through the model for variant genotyping, leading to greater accuracy in both variant calling and genotyping. DNAscope LR module further provides capability to process long reads as well.

With advancements in genomic data processing algorithms and a highly optimized implementation, DNAscope and DNAscope LR pipelines offer alignment, pre-processing, QC, variant calling and filtering functions, delivering results 5 to 10 times faster than the best open-source pipeline, offering both speed and precision in variant analysis.

About OmniTier Insight

There are 2 variants of Insight tertiary analysis:

1. Insight Oncology
2. Insight Hereditary

Insight Oncology is a tertiary analysis software tool that performs AI driven variant interpretation on either tumor or tumor+normal variant data, from gene panels to WES or WGS samples. Outputs include variant oncogenicity classification and reports containing AMP classification with associated drugs and therapies. For tumor+normal analysis, detection of hereditary cancer related variants is also supported.

Insight Hereditary is germline, hereditary and rare disease tertiary analysis software that runs on OmniTier’s CompStor Server platform, either on-premise, in an HPC or in a private cloud. Insight Hereditary can analyze single, duo, trio or larger family collections of gene panels, WES or WGS. Outputs can either be draft or final causal reports of prioritized, classified, pathogenic and likely pathogenic variants. Alternatively, cohorts of samples can be analyzed using genome-wide associate study features.

About OmniTier CompStor Workgroup Server

The OmniTier high-performance 1U rack server was designed for demanding workloads in modern data centers. Powered by the latest Intel Xeon processors, it offers exceptional scalability, storage options, and advanced management features. Ideal for virtualization, database applications, and cloud services, it ensures reliability and efficiency for enterprise operations. The CompStor Workgroup Server is scalable in configurations of 1, 2, 4, or 8 nodes per compute cluster. Cluster nodes communicate with each other with a 10 GB/s network; the customer network can be connected at 1-10 GB/s speeds to transfer workflow input and output files and allow access to the control GUI.

Key server features

• 1U rack mount server
• Intel Xeon Gold 16 Cores, 3rd generation, 2 sockets
• 512 GB DRAM
• 12TB NVMe SSD
• 2x 10GBase-T network ports.

Methods

Secondary Analysis using DNAscope

The correct version of DNAscope (version 202308.03) was downloaded as a binary from the Sentieon website and installed in a container on the OmniTier CompStor Workgroup server. A localhost trial license was issued to OmniTier for purposes of this benchmark testing. Sequencer specific scripts provided by Sentieon were then run for analysis purposes.

Tertiary Analysis using Insight

DNAscope generated VCF output files were then tested in Insight Hereditary.

Importing VCF Files: VCF files generated by DNAscope or any other secondary analysis tool can be imported through YML/CSV files.

Analysis Workflow: After importing a VCF data file into CompStor Insight, analysis of samples can be performed including:

• Annotating raw variants with genetic traits and disorders, population allele frequencies, functional effect prediction and other useful information
• Filtering variants according to a comprehensive list of criteria
• Classifying variants according to ACMG guidelines
• Prioritizing variants based on significance

Loading Analysis Results: Once an analysis is completed, users can load the results by clicking on the icons in the "Action" column. This will display a table containing all the analyzed variants.

Variant Details Page: To review the details of each variant, right-click on any variant row in the variant table to bring up the Single Variant Details Page. These details include allele frequency, functional effects, ACMG classification, a genome viewer (that automatically zooms to the selected variant).

Genome in a Bottle (GIAB) Test Data

The Genome in a Bottle Consortium is a public-private-academic consortium hosted by NIST to develop the technical infrastructure (reference standards, reference methods, and reference data) to enable translation of whole human genome sequencing to clinical practice and innovations in technologies.

The GIAB HG002 genome is a high-quality human reference genome derived from a well-characterized individual from the Ashkenazi Jewish population, part of the Ashkenazi Trio (HG002, HG003, HG004). It serves as a gold standard for evaluating and benchmarking genome sequencing technologies, variant calling methods, and bioinformatics pipelines.

HG002 versions were used for benchmarking that had been sequenced with Illumina, PacBio, and Oxford Nanopore sequencers, to get a range of overall and independent runtimes, along with secondary analysis F1 variant calling quality scores.

Results

Runtime results were measured for the end-to-end pipeline, from FASTQ or unaligned BAM input to DNAscope and through Insight tertiary analysis, to produce a draft report.

DNAscope secondary analysis performance:

Insight tertiary analysis performance:

Summary

Integration of the Sentieon DNAscope secondary analysis software into OmniTier’s CompStor Workgroup server enabled end-to-end analysis of Illumina, PacBio and Oxford Nanopore WGS data, on a single mid-range CPU based rack server. The speed optimization during the development of both DNAscope and Insight results in short end-to-end runtimes while keeping accuracy high. These good test results demonstrate that this is a good solution for those customers seeking an on-premise, end-to-end analysis solution.