Cross-platform comparisons for DNA methylation sequencing
Mingyang Su,Jin Sun,Jianhua Ma,Jun Xia,Chengcheng Ma,Wei Li,Zhixi Su,Qiye He,Rui Liu
DOI: https://doi.org/10.1158/1538-7445.am2022-6201
IF: 11.2
2022-01-01
Cancer Research
Abstract:Abstract Introduction: Accurate, reproducible, and cost-effective NGS platform is essential to identify malignancy-related abnormal DNA methylation changes and translate them into clinical applications including cancer detection, prognosis, and surveillance. In this study we evaluated the performances by the MGISEQ-2000 sequencer in DNA methylation sequencing, using Illumina’s NovaSeq6000 as a benchmark. Materials and Methods: We performed both genome-wide methylation profiling and targeted methylation sequencing on reference DNA and clinical DNA samples. To minimize non-platform discrepancies, we prepared 2 libraries, one for NovaSeq600 and the other for MGISEQ-2000, from each DNA sample using the same experimental protocol except the last step, when we used indexed primers whose sequences and chemical modification are compatible with either NovaSeq600 or MGISEQ-2000 to amplify the corresponding library. We conducted standardized analyses on MGISEQ-2000’s reads quality, sequence bias, genome alignment and coverage, regional depth, limits of detection (LOD), etc. We compared methylation levels at CpG sites, and measurements of methylation haplotypes to assess the consistency in mapping and quantifying genome-wide and regional methylation by MGISEQ-2000 and between the 2 platforms. Furthermore, we analyzed inter-platform variations in classifying normal and malignant plasma DNA samples using targeted methylation sequencing data from either platform exclusively. Results: We prepared and sequenced libraries from over 100 DNA samples of normal or malignant tissue or blood samples for comparison, in addition to reference DNA. Overall, MGISEQ-2000 displayed high level of consistency, and little-to-no sequence bias in methylation sequencing. At genome-wide, regional, and individual-CpG levels, MGISEQ-2000 demonstrated high degree of concordance with NovaSeq6000 in quantifying DNA methylation, as the inter-platform Pearson correlation coefficients were consistently above 0.90 or higher for both targeted sequencing and whole-genome bisulfite sequencing, regardless of whether average methylation level (AMF) or methylation haplotype-specific metrics such as Methylation Haplotype Load (MHL) were used to quantify methylation. Additionally, MGISEQ-2000 is as sensitive as NovaSeq6000 at detecting spik-in methylation controls, as both reached an LOD of 0.1%. Lastly, MGISEQ-2000 produced similar results in classifying normal and malignant samples as NovaSeq6000, suggesting it had minimal distortion on the NovaSeq6000-based classifier. Conclusions: MGISEQ-2000 demonstrated comparable data quality, methylation-calling accuracy, and consistency as NovaSeq6000 in DNA methylation sequencing, supporting its application in future investigation on DNA methylation as cancer biomarkers. Citation Format: Mingyang Su, Jin Sun, Jianhua Ma, Jun Xia, Chengcheng Ma, Wei Li, Zhixi Su, Qiye He, Rui Liu. Cross-platform comparisons for DNA methylation sequencing [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 6201.