Optimizing a massive parallel sequencing methodology for whole mitochondrial genome analysis at the diagnostic bench.

Kim Vancampenhout, Ben Caljon, Linda De Meirleir, Willy Lissens, Katrien Stouffs, An Jonckheere, Rudy Van Coster, Sara Seneca

Research output: Chapter in Book/Report/Conference proceedingMeeting abstract (Book)Research

Abstract

Lately, sequencing technologies have been evolving rapidly and at this moment it is possible to analyse a complete human (g)e(x)(n)ome sequence within days at an affordable cost. However, some of the next generation sequencing (NGS) methodologies come with a pitfall, and different biases of the resulting data can be generated. A well-known limitation is the nucleotide sequencing infidelity in the vicinity of homopolymer stretches. Other shortcomings involve coverage and therefore variant calling. The mitochondrial DNA (mtDNA) is a maternally inherited, small, double stranded circular molecule (16569 base pairs) encoding essential genetic information implicated in the energy production pathway of an eukaryotic cell. Numerous mtDNA molecules, which can be all identical (homoplasmy) or different (heteroplasmy) reside within a cell. In a recent study we re-sequenced the mitochondrial genome of 32 human samples using the Ion Torrent PGM sequencer system. Remarkably, coverage analysis of the resulting massive parallel sequencing data exposed a significant difference in read depth between the plus and the minus strand for about 10% of the regions in this small genome. The discrepancy was exclusively seen for the mitochondrial genome data and never for the sequencing results of pUC19 plasmid DNA samples. This strand bias may well hinder variant calling of low heteroplasmic nucleotide alterations. Indeed, identification of allele modifications might become challenging as the differentiation between true minority allele frequencies and sequencing errors was small. To overcome this limitation, we envisaged a different shearing approach to reduce or even eliminate the strand bias of our initial Ion Torrent PGM data, and mechanical and enzymatic fragmentation methodologies of the mitochondrial genome were compared. Six DNA samples were sheared both with the Covaris M220 sonificator, and the Nebnext dsDNA fragmentase. All sheared DNAs were further processed with standard Ion Torrent protocols. In addition, these sheared samples were also used in a TrueSeq library preparation protocol, and subsequently sequenced on an Illumina HiSeq1500 device. The use of alternative shearing methods within the Ion Torrent PGM system did not at all ameliorate the strand bias problematic for the mitochondrial genome. However, all Illumina results were in strong contrast to these data. A uniform coverage of both strands was observed in relation to the mtDNA sequences. We therefore hypothesize that the troughs and peaks generated by the Ion Torrent PGM system rather originate from the proliferation of the sheared mtDNA sequences and not from the fragmentation method in se. In fact, it might be characteristic to the combination of the nature of the mitochondrial genome sequence configuration on the one hand and the library PCR amplification step included in the standard protocol on the other hand.
Original languageEnglish
Title of host publicationBeSHG
Pages36-36
Number of pages1
Publication statusPublished - 7 Feb 2014

Publication series

Name
NumberT5

Keywords

  • NGS
  • MPS
  • mtDNA

Fingerprint

Dive into the research topics of 'Optimizing a massive parallel sequencing methodology for whole mitochondrial genome analysis at the diagnostic bench.'. Together they form a unique fingerprint.

Cite this