Communicated by Mats Nilsson
SNP discovery performance of two second-generation sequencing platforms in the NOD2 gene region†
Article first published online: 18 MAY 2010
© 2010 Wiley-Liss, Inc.
Volume 31, Issue 7, pages 875–885, July 2010
How to Cite
Melum, E., May, S., Schilhabel, M. B., Thomsen, I., Karlsen, T. H., Rosenstiel, P., Schreiber, S. and Franke, A. (2010), SNP discovery performance of two second-generation sequencing platforms in the NOD2 gene region. Hum. Mutat., 31: 875–885. doi: 10.1002/humu.21276
- Issue published online: 24 JUN 2010
- Article first published online: 18 MAY 2010
- Manuscript Accepted: 19 APR 2010
- Manuscript Received: 23 NOV 2009
- second-generation sequencing;
- rare variants;
- coverage simulation;
- SNP discovery
A potentially important application of second generation sequencing technologies is to identify disease-associated variation. For comparison of the performance in SNP detection, the Crohn's disease (CD)-associated NOD2 gene was subjected to targeted resequencing using two different second-generation sequencing technologies. Eleven CD patients were selected based on their haplotype background at the NOD2 locus. The 40-kb large NOD2 gene region was amplified using long-range PCR (LR-PCR), and sequenced with the Roche 454/FLX system, an Applied Biosystems SOLiD mate-pair library (2×25 bp), and a SOLiD fragment (50 bp) library. The entire NOD2 region was also sequenced using conventional Sanger technology. Four-hundred forty-two single nucleotide polymorphisms (SNPs) were discovered with the SOLiD mate-pair library, 454 with the fragment library, and 441 with the 454/FLX. For the homozygous SNPs, 98% were confirmed by Sanger for the mate-pair library, 100% for the fragment library and 99% for the 454/FLX. Ninety-six percent of the heterozygous SNPs detected with the SOLiD mate-pair library, 91% with the fragment library and 96% with the 454/FLX were confirmed by Sanger. In a simulation, the SNP detection performance fell rapidly when the achieved coverage was below 40×. Due to uneven representation of the target region when using LR-PCR, oversequencing of other regions is necessary. Hum Mutat 31:875–885, 2010. © 2010 Wiley-Liss, Inc.