Analysis of genetic variants in myeloproliferative neoplasms using a 22-gene next-generation sequencing panel

Abstract

Background: The Philadelphia (Ph)-negative myeloproliferative neoplasms (MPNs), namely essential thrombocythaemia (ET), polycythaemia vera (PV) and primary myelofibrosis (PMF), are a group of chronic clonal haematopoietic disorders that have the propensity to advance into bone marrow failure or acute myeloid leukaemia; often resulting in fatality. Although driver mutations have been identified in these MPNs, subtype-specific markers of the disease have yet to be discovered. Next-generation sequencing (NGS) technology can potentially improve the clinical management of MPNs by allowing for the simultaneous screening of many disease-associated genes.

Methods: The performance of a custom, in-house designed 22-gene NGS panel was technically validated using reference standards across two independent replicate runs. The panel was subsequently used to screen a total of 10 clinical MPN samples (ET n = 3, PV n = 3, PMF n = 4). The resulting NGS data was then analysed via a bioinformatics pipeline.

Results: The custom NGS panel had a detection limit of 1% variant allele frequency (VAF). A total of 20 unique variants with VAFs above 5% (4 of which were putatively novel variants with potential biological significance) and one pathogenic variant with a VAF of between 1 and 5% were identified across all of the clinical MPN samples. All single nucleotide variants with VAFs ≥ 15% were confirmed via Sanger sequencing.

Conclusions: The high fidelity of the NGS analysis and the identification of known and novel variants in this study cohort support its potential clinical utility in the management of MPNs. However, further optimisation is needed to avoid false negatives in regions with low sequencing coverage, especially for the detection of driver mutations in MPL.

Keywords: Bioinformatics, disease management; Gene; Mutation; Myeloproliferative neoplasm; Next-generation sequencing; Variant.

Fig. 1

Schematic representation of steps involved in the quality control and bioinformatics analysis of the NGS data. The variant filtering and prioritisation step is further illustrated in Fig. 3. Adapted from Dai et al., 2019 [11] and Zheng et al., 2018 [12]

Fig. 2

Combined analysis of variants with the DNA Amplicon and Pindel apps to evaluate the performance of the custom 22-gene NGS panel based on two identical but independent NGS runs using reference standards. The MPL W515L variant in the Seraseq Myeloid Mutation Mix (Seraseq) was not detected in one of the replicates, giving the custom NGS panel a sensitivity of 99.2%. One variant was detected in the wild-type reference standard TruQ0, giving the panel a specificity of 96.3%. The custom NGS panel also had a positive predictive value of 97.7%, an average intra-run and inter-run concordance of 98.8% [range 95.2–100%] and 99.0% [range 95.2–100%] respectively, and was able to detect variants at as low as 1% allele frequency. FP, False positive; TP, True positive; FN; False negative; Rep 1, Replicate 1; Rep 2, Replicate 2

Fig. 3

Variant filtering and prioritization process. Note that after this process, variants with VAFs between 1 and 5% were inspected for the presence of any pathogenic/likely pathogenic variants. Adapted from Zheng et al., 2018 [12]

Fig. 4

Variants detected in the clinical MPN samples with VAF > 5%. Note that Sample 02 also carries another variant in TET2 with a VAF of 1.4% (not shown in Fig. 4)