Next generation sequencing for clinical diagnostics-principles and application to targeted resequencing for hypertrophic cardiomyopathy: a paper from the 2009 William Beaumont Hospital Symposium on Molecular Pathology

Abstract

During the past five years, new high-throughput DNA sequencing technologies have emerged; these technologies are collectively referred to as next generation sequencing (NGS). By virtue of sequencing clonally amplified DNA templates or single DNA molecules in a massively parallel fashion in a flow cell, NGS provides both qualitative and quantitative sequence data. This combination of information has made NGS the technology of choice for complex genetic analyses that were previously either technically infeasible or cost prohibitive. As a result, NGS has had a fundamental and broad impact on many facets of biomedical research. In contrast, the dissemination of NGS into the clinical diagnostic realm is in its early stages. Though NGS is powerful and can be envisioned to have multiple applications in clinical diagnostics, the technology is currently complex. Successful adoption of NGS into the clinical laboratory will require expertise in both molecular biology techniques and bioinformatics. The current report presents principles that underlie NGS including sequencing library preparation, sequencing chemistries, and an introduction to NGS data analysis. These concepts are subsequently further illustrated by showing representative results from a case study using NGS for targeted resequencing of genes implicated in hypertrophic cardiomyopathy.

Figure 1

Next generation sequencing process steps for platforms requiring clonally amplified templates (Roche 454, Illumina and Life Technologies). Input DNA is converted to a sequencing library by fragmentation, end repair, and ligation to platform specific oligonucleotide adapters. Individual library fragments are clonally amplified by either (1) water in oil bead–based emulsion PCR (Roche 454 and Life Technologies) or (2) solid surface bridge amplification (Illumina). Flow cell sequencing of clonal templates generates luminescent or fluorescent images that are algorithmically processed into sequence reads.

Figure 2

A: Coverage plot of a region of the MYH7 gene with comparison of Roche 454 GS FLX (top panel) and Illumina Genome Analyzer (bottom panel) sequence read data. Coverage is shown on the y axis and MYH7 reference sequence position on the x axis. Overlapping long-range PCR amplicons and exon positions are shown. Note difference in coverage scales between platforms. B: Variant read percentage plot of a region of the MYH7 gene showing variants identified with the Roche 454 GS FLX and Illumina Genome Analyzer platforms. Variant read percentage is shown on the y axis and MYH7 reference sequence position on the x axis. Exons and introns are indicated. Labeled as heterozygote and homozygote are two platform-concordant Sanger-confirmed variants with read percentages consistent with heterozygosity and homozygosity, respectively. Also circled are three variants with read percentages of approximately 25%. Of these three variants, the two identified only with the Illumina Genome Analyzer were determined to be false positives by Sanger sequencing. The third variant, with a read percentage of approximately 25% with the Roche 454 GS FLX, was determined to be a true variant. The complement of this variant is present in the Illumina Genome Analyzer data at a read percentage of approximately 50% (black circle with asterisk).

Figure 3

Comparison examples of Roche 454 GS FLX and Illumina Genome Analyzer next generation sequencing results for hypertrophic cardiomyoapathy associated genes. A: Concordant heterozygous variant g.5871488T>C in ACTC1. Roche 454 GS FLX and Illumina Genome Analyzer sequence read data were aligned and viewed in SeqMan NGen version 1.2 software (DNAstar, Madison, WI). Screen shots are shown. Top: GS FLX data showing 19 of 84 reads with the ACTC1 reference sequence shown above the panel box and the variant base designated in blue. Variant read percentage for 84 reads is 50%. Middle: Genome Analyzer data showing 21 of 499 reads and variant base in blue. Variant read percentage for 499 reads is 51.7%. Bottom: Confirmatory Sanger electropherogram. B: Discordant exon variant g.11968221delC in PRKAG2 due to GS FLX homopolymer sequencing error. Roche 454 GS FLX and Illumina Genome Analyzer sequence read data were aligned and viewed in SeqMan NGen version 1.2 software (DNAstar, Madison, WI). Screen shots are shown. Top: GS FLX data showing 19 of 37 reads with the PRKAG2 reference sequence, containing a 7-base polyC tract, shown above the panel box. The g.11968221delC variant in the 7-base poly C tract is present in 32.4% of 37 reads and designated with a dash mark. Middle: Genome Analyzer data showing 19 of 279 reads. The g.11968221delC variant is present in 1.8% of 278 reads. Bottom: Confirmatory Sanger electropherogram. C: Discordant exon variant g.4892783C>T in MYH7 due to misalignment of Genome Analyzer 36-base length reads generated from MYH6 in pooled amplicon library. Roche 454 GS FLX and Illumina Genome Analyzer sequence read data were aligned and viewed in SeqMan NGen version 1.2 software (DNAstar, Madison, WI). Screen shots are shown. Top: GS FLX data showing 19 of 39 reads with the MYH7 reference sequence shown above the panel box. All 39 reads show reference C at position g.4892783. Middle: Genome Analyzer data showing 19 of 293 reads and the variant base designated in blue. Variant read percentage for 293 reads is 21.8%. Bottom: Confirmatory Sanger electropherograms for MYH7 (upper trace) and MYH6 (lower trace) showing variant g.4858071C>T. In this region, MYH7 and MYH6 are 100% identical over 59 bases in the respective reference sequences. Genome Analyzer reads containing the MYH6 variant g.4858071C>T cross aligned to the MHY7 reference sequence in this region resulting in the discordant variant. The longer GS-FLX reads extend beyond this local region of MYH6 and MYH7 homology and avoided misalignment. An additional feature, highlighted in yellow, is a single G/A base difference between MYH6 and MYH7 downstream from the discordant variant. MYH6 Genome Analyzer reads overlying this G/A base difference aligned to MYH6 and not to MYH7.