Abstract 220: Performance assessment of the novel G4 sequencing platform for cancer research applications

Abstract

Abstract Introduction: Next-generation sequencing (NGS) has become a critical tool for both biological research and in-vitro diagnostics, though there remains a need for new DNA sequencing platforms that combine high accuracy, speed, and flexible throughput to provide timely results and cost-effective operations. Here we evaluate the utility of the novel Singular Genomics G4 sequencing platform for cancer research applications by performing whole exome sequencing (WGS) of the human reference control HG001. We further demonstrate performance for methyl-seq, bulk RNA-seq, and scRNA-seq. Methods: To assess performance, we prepared a human whole exome library (IDT xGen Exome) from HG001 gDNA, then sequenced the library via the G4 sequencing platform with F2 flowcells (150M read throughput) to yield 483M 2x150bp reads. Sequence quality and mapping metrics were obtained via GATK4 after read alignment to Grch38 via bwa mem. Germline variants were identified via DeepVariant v1.4 using the Illumina WES model, implemented via Parabricks; performance was assessed by hap.py. Methyl-seq performance was examined by methylDackel analysis of an EM-Seq library (NEB) prepared from 50ng HG001 gDNA and sequencing to 20x coverage via 2x150bp reads. Bulk RNA-seq performance was assessed by replicate sequencing of UHR total RNA on the G4 and Illumina NextSeq platforms. Single cell RNA-seq libraries were prepared from healthy donor PBMC via the 10x Chromium, sequenced on the G4 and Illumina NovaSeq, then analyzed via Cell Ranger, Scanpy and scvi-tools. Results: Exome sequencing of HG001 yielded high quality data (single pass accuracy > 99.7% for R1 and R2) with even coverage of the genome (fold 80 =1.42). hap.py analysis of DeepVariant calls yielded a SNP and Indel F1 of 99.06 and 96.05%, respectively, over high confidence regions. Performance was high over challenging features such as low complexity DNA and GC rich regions. EM-Seq analysis revealed global distributions of cytosine methylation that closely matched public EM-Seq data derived from the Illumina NovaSeq. Bulk RNA-seq transcript counts were highly correlated across technical replicates (R2 = .990) and platforms (R2= .988). Pseudo-bulk analysis of scRNA-seq libraries showed high cross-platform correlation (R2= .982), with nearly identical CellTypist annotations (Adjusted Rand Index = .999). Conclusions: The G4 sequencer delivers sequencing data on par with the state-of-the-art in NGS, but with a faster turnaround than traditional SBS systems. Notably, we observe a high concordance with the leading NGS platform for human DNA variant detection, methyl-seq, bulk RNA-seq and scRNA-Seq. We anticipate the rapid turnaround and flexible throughput of the G4 will have broad utility for basic and translational cancer research applications. Citation Format: Kenneth Gouin, Liz LaMarca, Yu Xiang, Anne Decker, Sabrina Shore, Ryan Shultzaberger, Daan Witters, Timothy Looney, Martin M. Fabani, Eli Glezer. Performance assessment of the novel G4 sequencing platform for cancer research applications [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 220.