SAN JOSE, Calif., April 29, 2026 /PRNewswire/ -- Complete Genomics, a U.S.-based leader in genomic sequencing technologies, today announced results demonstrating that AI-trained variant calling models can reduce genome analysis errors by up to 73% when applied to DNBSEQ™ sequencing data generated on its advanced massively parallel sequencing platform compared to widely used industry standard pipelines.

Accurate genome interpretation is essential for applications ranging from early disease detection to targeted therapies. However, traditional analysis methods can introduce errors that limit reliability and increase downstream validation requirements. The results from this study show how integrating high-quality sequencing with AI models trained specifically on DNBSEQ data can significantly improve accuracy and consistency across the genome, including in complex and medically relevant regions.

"This is about making high-quality genomic analysis scalable and economically viable," said Radoje Drmanac, Ph.D., Founder and Chief Scientific Officer of Complete Genomics. "By integrating AI directly with our PCR-free library prep and clonal-error-free sequencing platforms, we are enabling more accurate detection of genetic variation, including in regions that have traditionally been difficult to analyze."

"Across multiple datasets generated from the T1+, T7 and T7+ platforms, we observed consistently high accuracy for SNVs and indels using DeepVariant models trained on DNBSEQ data. These results highlight the robustness of the approach for scalable variant calling," said Andrew Carroll, Product Lead at Google Research.

The study also showed that the platform can identify genetic changes in DNA more accurately and with fewer errors than commonly used methods. This also includes more complex sequence alterations, with performance on these difficult genomic regions approaching that of long-read sequencing technologies.

Together, these findings highlight the potential of combining advanced sequencing platforms with novel AI-driven analysis to deliver high-accuracy genomics at scale. As genomic data continues to expand in importance across healthcare and life sciences, improvements in accuracy, efficiency, and scalability will be key to enabling broader adoption in both research and translational settings.

PanVariants, the AI-driven analysis framework used in the study, is being made available as an open-source resource to support further innovation across the genomics community.

Complete Genomics will host a webinar on April 30 to further explore how high-throughput sequencing and AI-driven variant calling are advancing genomic analysis.

To register, visit:
https://www.completegenomics.com/webinars/webinar-scaling-genomics-with-higher-throughput-and-ai-driven-variant-calling

About Complete Genomics

Complete Genomics is a U.S.-based life sciences company providing end-to-end DNA sequencing platforms, reagents, and software solutions. Since its founding in 2005, the company has contributed to more than 10,900 scientific publications and continues to advance high-throughput, cost-effective genomics.

*For Research Use Only. Not for use in diagnostic procedures.

View original content:https://www.prnewswire.com/news-releases/new-study-demonstrates-up-to-73-reduction-in-genome-analysis-errors-using-ai-trained-models-versus-standard-approaches-302757446.html

SOURCE Complete Genomics