Fig. 2

MGISEQ-2000 showed similar data quality with NovaSeq6000. A Design of cross-platform comparison. In brief, we compared the targeted methylation sequencing of NovaSeq6000 and MGISEQ-2000 with the synthetic cfDNA samples and clinical cfDNA samples. The synthetic cfDNA samples were made by spiking pancreatic ductal adenocarcinoma (PDAC) gDNA into NA12878 at tumor fractions of 0, 0.1%, 0.5%, 1%, 5%, and 10% (four replicates for each tumor fraction). We also used 24 cfDNA samples, which were from 12 PDAC patients and 12 healthy donors. Two libraries were prepared with Illumina official experimental kits and MGI official experimental kits, and renamed as “iLib” and “mLib”. The iLib were sequenced by NovaSeq6000 and mLib were sequenced by NovaSeq6000 and MGISEQ-2000, which were finally allocated to three data ‘iLib-NovaSeq’, ‘mLib-NovaSeq’ and ‘mLib-MGISEQ’. The analysis pipeline was shown on the right panel. B Boxplot plot showing the sequencing error rate of iLib-NovaSeq, mLib-NovaSeq, and mLib-MGISEQ data. The statistical analysis was performed by “Wilcox. test” and adjusted by "holm". C The distribution of insert size of sequencing data of the synthetic cfDNA samples. The x-axis represented the insert size of alignments, the y-axis represented the density of insert size distribution; colors represented data types. D The percentage of alignments in different insert size intervals. We made 50 bp-bin intervals and summarize the percentage of alignments in the intervals. Filled colors represented the intervals of insert size