Analytical evaluation of whole genome sequencing for acute myeloid leukemia

BackgroundAcute myeloid leukemia (AML) is the most common leukemia in adults and current methods rely on cytogenic and molecular profiling for AML classification. Recent advances have demonstrated that next generation sequencing (NGS) may improve prognostic prediction and risk stratification of AML patients. A tumour-only high coverage whole genome sequencing (WGS) method was developed to assess the analytical performance (limit of detection (LoD)), sensitivity, specificity and reproducibility) in identifying somatic small variants, structural variants (SV) and copy number alterations (CNA) specific to AML.MethodsSeracare® myeloid panel, NOMO1, Kasumi1, HCC1187 and 53 AML clinical specimens were used to assess analytical performance. Extracted genomic DNA (∼350ng) from contrived specimens, cell lines, mononuclear cells from whole blood or bone marrow aspirate were manually processed with Illumina DNA PCR Free Prep and sequenced at 4 plex on S4 flowcells using NovaSeq™ 6000. The sequence data was analysed by DRAGEN™ Somatic v4.2 with a custom virtual 156 AML specific gene panel variant classification algorithm.ResultsThe samples were sequenced at an average of ∼220X which yielded high analytical sensitivity, accuracy and reproducibility of the specimens evaluated in this study. Although the target sequence was ∼220X, samples were in silico down sampled to 140X to establish the following performance specifications. LoD was established for small variant (SNVs and indels) and SV to be 5% and 7.3%, respectively. CNA were detected with copy number fold change of 1.09 and 0.87 for duplication and deletion events, respectively. Further, loss of heterozygosity was detected with purity of 17%. Overall, we observed accuracy >95% across clinical samples with analytical sensitivities >95%, >99% and >95% for small variants, SV and CNA, respectively and analytical specificity of >99.9%.ConclusionsWe have developed a tumour only WGS method for genomic profiling of AML patients with a turn around time of 5 days. The analytical performance of this tumour only WGS approach indicates high accuracy in identifying variants indicative for AML. In addition, the high sequence depth provides improved analytical sensitivity compared to other genome sequencing or cytogenetic methods.Legal entity responsible for the studyIllumina.FundingIllumina.DisclosureG. Malagoli Tagliazucchi, W. Gong, M. Ghildiyal, M. Chavez, K. Nobuta, E. Sayyari, S. Truong, T. Vang, G. Kim, Q. Bui, S. Kim, J. Bernd, E. De Feo, S. Comer: Financial Interests, Institutional, Full or part-time Employment: Illumina. M. Yaylaoglu: Financial Interests, Personal and Institutional, Full or part-time Employment: Illumina. D. Spencer: Financial Interests, Institutional, Financially compensated role: Washington University in St. Louis.