Detection of mutations in myeloid malignancies through paired-sample analysis of Microdroplet-PCR deep sequencing data Academic Article uri icon

Overview

MeSH Major

  • High-Throughput Nucleotide Sequencing
  • Leukemia, Myeloid
  • Mutation
  • Myeloproliferative Disorders
  • Polymerase Chain Reaction

abstract

  • Amplicon-based methods for targeted resequencing of cancer genes have gained traction in the clinic as a strategy for molecular diagnostic testing. An 847-amplicon panel was designed with the RainDance DeepSeq system, covering most exons of 28 genes relevant to acute myeloid leukemia and myeloproliferative neoplasms. We developed a paired-sample analysis pipeline for variant calling and sought to assess its sensitivity and specificity relative to a set of samples with previously identified mutations. Thirty samples with known mutations in JAK2, NPM1, DNMT3A, MPL, IDH1, IDH2, CEBPA, and FLT3, were profiled and sequenced to high depth. Variant calling using an unmatched Hapmap DNA control removed a substantial number of artifactual calls regardless of algorithm used or variant class. The removed calls were nonunique, had lower variant frequencies, and tended to recur in multiple unrelated samples. Analysis of sample replicates revealed that reproducible calls had distinctly higher variant allele depths and frequencies compared to nonreproducible calls. On the basis of these differences, filters on variant frequency were chosen to select for reproducible calls. The analysis pipeline successfully retrieved the associated known variant in all tested samples and uncovered additional mutations in some samples corresponding to well-characterized hotspot mutations in acute myeloid leukemia. We have developed a paired-sample analysis pipeline capable of robust identification of mutations from microdroplet-PCR sequencing data with high sensitivity and specificity.

publication date

  • January 2014

Research

keywords

  • Academic Article

Identity

Language

  • eng

Digital Object Identifier (DOI)

  • 10.1016/j.jmoldx.2014.05.006

PubMed ID

  • 25017477

Additional Document Info

start page

  • 504

end page

  • 18

volume

  • 16

number

  • 5