Overview of model training and benchmarking pipeline.

DNAscope, a germline variant-calling pipeline developed by San Jose–based analytics firm Sentieon, has demonstrated superior SNP and indel accuracy compared with standard variant-calling datasets, according to a recent company preprint. In the study, DNAscope was paired with MGI StandardMPS chemistry, with a new DNAscope model trained specifically for the MGI DNBSEQ-G400* sequencing platform.

The Genome Analysis Toolkit (GATK) HaplotypeCaller is widely regarded as the industry standard for small variant calling. However, existing short-read variant callers, including HaplotypeCaller, still show imperfect concordance with high-confidence variant calls, particularly in clinically relevant, complex genomic regions. As next-generation sequencing sees increasing clinical adoption, improving accuracy in these regions is becoming more critical, the study authors noted.

Sentieon’s DNAscope combines established haplotype-based variant-calling methods with machine learning to improve accuracy. The pipeline enhances active region detection and local assembly, delivering greater sensitivity and robustness, especially in high-complexity regions. DNAscope outputs candidate variants with informative annotations, which are then processed by a machine-learning model for variant genotyping.

DNAscope MGI model training on DNBSEQ-G400

The researchers developed a DNAscope MGI model to evaluate the accuracy of PE150 whole genome sequencing data generated on MGI’s DNBSEQ-G400 medium-throughput benchtop sequencer. The platform was selected for its flexibility and comprehensive performance across a range of sequencing applications.

When combined with the trained DNAscope model, the DNBSEQ-G400 delivered higher accuracy than previously published benchmarks and other mainstream platforms. The workflow also achieved faster processing speeds, while reducing both false negative and false positive variant calls. This improvement was attributed to DNAscope’s ability to model systematic sequencing error patterns more effectively.

“MGI provides high-quality sequencing products, giving the genomics community more choice in sequencing platforms,” said Jun Ye, CEO of Sentieon. “We are very pleased to optimise DNAscope for MGI sequencers and jointly launch this secondary analysis solution for MGI customers. This optimised pipeline delivers high accuracy, computing efficiency, and ease of use. We look forward to continued collaboration to provide high-quality solutions for the genomics industry.”

With its combination of high accuracy and low cost, the DNBSEQ-G400 has the potential to make high-performance sequencing projects more accessible to researchers and clinicians, particularly those working on human disease and population diversity using large sample cohorts. When paired with Sentieon’s DNAscope, the system enables improved variant calling in complex, clinically relevant genomic regions, supporting more meaningful clinical insights and treatment strategies.

As demonstrated in the study, the DNBSEQ-G400 optimises day-to-day sequencing while delivering higher SNP and indel detection accuracy, according to Yongwei Zhang, CEO of MGI Americas. The system offers a daily data output of up to 1440 Gb, with support for one or two flow cells, two flow cell types (550M reads or 1800M reads), and multiple read length options ranging from SE50 to SE400 or PE300.

Built on a new flow cell system with optimised optical and biochemical designs, the DNBSEQ-G400 supports a wide range of sequencing and data analysis applications, including basic research, clinical research, forensics, and agriculture. Following its US release in August, the first systems were installed in customer laboratories within a week, reflecting MGI’s commitment to local customers and partners.

“MGI competes on the strength of our DNBSEQ technologies, complete workflow solutions covering upstream automation and downstream BioIT, strong local service teams, and cost,” Zhang said. “Our goal is to give customers more choice through higher data quality and lower costs, enabling cutting-edge research and more precise, affordable clinical testing.”