B-214 Migration of a Previously Validated Sequencing Workflow of 4,700 Genes Associated with Clinical Phenotypes to the MGI DNBSEQ-G400 Platform: High-Quality NGS Data with Elevated Concordance to Expected Results

Background Advances in next-generation sequencing (NGS) have significantly enhanced clinical diagnostics by enabling comprehensive analysis of genes associated with various inherited disorders. This study aimed to transfer a previously validated sequencing workflow—originally executed on the NextSeq-500 (Illumina, Inc.)—to the DNBSEQ-G400 platform (MGI Tech Co.). The workflow targets a panel of 4,700 genes with well-established gene-disease associations that span a broad range of clinical phenotypes. Our primary objectives were to evaluate the performance of the DNBSEQ-G400 in terms of data yield, quality, and variant calling accuracy, and to assess its concordance with variants (SNVs and INDELs) detected on the NextSeq-500. Methods An NGS library representing 12 patients—whose saliva-derived DNA was prepared using the xGEN DNA Library Prep kit and enriched with the xGen Inherited Disease Panel V1 (covering 4,700 genes) and Hybridization Capture kit (all from Integrated DNA Technologies, Inc.)—was initially sequenced on the NextSeq-500 platform (paired-end 150-cycle runs) and also on the DNBSEQ-G400 platform (paired-end 300-cycle runs). For the DNBSEQ-G400, an additional step was required: the libraries were processed with the DNBSEQ™ OneStep DNB Make Reagent Kit V4.0 to convert the linear DNA into DNA nanoballs. Data processing, including sequence alignment and variant calling, was performed using the Dragen v.4.2.4 pipeline on the EMEDGENE platform (Illumina, Inc.). For each sample, key performance metrics for the DNBSEQ-G400 were computed—namely, the percentage of bases with a quality score > Q30, total read counts, percentage of mapped reads, average coverage depth, percentage of bases with coverage >20X, and duplication rate. In addition, the study compared the number of single nucleotide variants (SNVs) and insertion/deletion events (INDELs) and calculated the F-measure—a metric integrating precision and recall—between the DNBSEQ-G400 and the NextSeq-500 platforms. Data are presented as medians with the corresponding minimum and maximum ranges. Results The DNBSEQ-G400 generated a median of 54,143,871 reads per sample (range: 31,912,278–72,109,870). A median of 93.06% of bases exceeded a Q30 score (range: 93.03%–93.22%), and 99.1% of reads mapped to the human reference genome (range: 98.02%–99.32%). The median average coverage was 397X (range: 230X–537X), the median percentage of bases with coverage >20X was 99.7% (range: 99.60%–99.80%), and the median duplication rate was 13.52% (range: 12.37%–15.09%). Variant calling identified a median of 7,709 SNVs (range: 7,437–8,028) and 344 INDELs (range: 329–367) per sample. The comparative F-measure was 99% for SNVs (range: 98.5%–99.5%) and 82.3% for INDELs (range: 79.3%–84.6%). Conclusion The results suggest that the workflow can be migrated from the NextSeq-500 to the MGI DNBSEQ-G400 platform. The DNBSEQ-G400 generated high-quality NGS data, as evidenced by robust base quality (median 93.06% of bases > Q30), high mapping rates (median 99.1%), and strong SNV detection (F-measure of 99%). Although the F-measure for INDELs was 82.3%, this is expected given the inherent challenges in indel detection. A limitation of this study is the difference in sequencing cycle numbers (paired-end 150-cycle vs. 300-cycle runs), which may affect direct comparisons. These findings support the adoption of the DNBSEQ-G400 platform in clinical settings.