Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2010 Jun;38(10):e116.
doi: 10.1093/nar/gkq072. Epub 2010 Feb 17.

Accurate SNP and mutation detection by targeted custom microarray-based genomic enrichment of short-fragment sequencing libraries

Michal Mokry  1 , Harma Feitsma, Isaac J Nijman, Ewart de Bruijn, Pieter J van der Zaag, Victor Guryev, Edwin Cuppen
Affiliations

Accurate SNP and mutation detection by targeted custom microarray-based genomic enrichment of short-fragment sequencing libraries

Michal Mokry et al. Nucleic Acids Res. 2010 Jun.

Abstract

Microarray-based enrichment of selected genomic loci is a powerful method for genome complexity reduction for next-generation sequencing. Since the vast majority of exons in vertebrate genomes are smaller than 150 nt, we explored the use of short fragment libraries (85-110 bp) to achieve higher enrichment specificity by reducing carryover and adverse effects of flanking intronic sequences. High enrichment specificity (60-75%) was obtained with a relative even base coverage. Up to 98% of the target-sequence was covered more than 20x at an average coverage depth of about 200x. To verify the accuracy of SNP/mutation detection, we evaluated 384 known non-reference SNPs in the targeted regions. At approximately 200x average sequence coverage, we were able to survey 96.4% of 1.69 Mb of genomic sequence with only 4.2% false negative calls, mostly due to low coverage. Using the same settings, a total of 1197 novel candidate variants were detected. Verification experiments revealed only eight false positive calls, indicating an overall false positive rate of less than 1 per approximately 200,000 bp. Taken together, short fragment libraries provide highly efficient and flexible enrichment of exonic targets and yield relatively even base coverage, which facilitates accurate SNP and mutation detection. Raw sequencing data, alignment files and called SNPs have been submitted into GEO database http://www.ncbi.nlm.nih.gov/geo/ with accession number GSE18542.

PubMed Disclaimer

Figures

Figure 1.
Figure 1.
Size distribution of human exons. The median size of human exons is only 120 bp (with 70% of all exons shorter than 200 bp). Therefore many of the specifically captured DNA from long-fragment libraries will consist of sequences derived from introns flanking the exons of interest, which decreases the effective sequencing yield.
Figure 2.
Figure 2.
Comparison of sequence coverage evenness after enrichment. The fraction of target positions with at least that coverage was as compared to the average coverage. (A) Comparison of various enrichment (washing temperature and input DNA) and sequencing (35- versus 50-mer) conditions. Library 1 sequenced by 50-mer reads results in the most even coverage compared to other libraries. The brown curve depicts the best possible evenness for an ideal evenly enriched sample with 100 × average coverage, where the unevenness is purely caused by statistical randomness in the coverage assuming a Poisson distribution of the sequencing reads. (B) The evenness score, E, represents the fraction of whole sequencing throughput that is correctly distributed (marked area below the curve). Consequently, 1-E represents the fraction of the (whole) sequencing output that has to be redistributed from positions with coverage above average to positions with coverage below average (by better enrichment) to get the ideal even coverage over all targeted positions. The more even the coverage, the higher the evenness score. (C) Correlation of evenness score E for randomized sets to the sequencing depth. In this simulation, the unevenness of these datasets is purely caused by the random distribution of reads and fits a Poisson distribution of sequence coverage. When the discrete character of the data is reduced by sufficient depth of coverage, E changes only slightly with increasing average coverage and thus can be characterized as relatively independent of sequencing depth. (D) Comparison between + and – strand coverage for 35- and 50-mer reads. In the case of 50-mer reads, the coverage is more even with fewer positions covered by extremely low (or high) numbers of sequencing tags. This difference is more prominent when the coverage is determined separately for the positive or negative strand. Independent strand coverage is better for 50-mer than for 35-mer sequencing.
Figure 3.
Figure 3.
Exemplary representation of target coverage after enrichment. Sequencing results of Library 1 are shown for 35-mer (green) and 50-mer (purple) sequencing. Total, positive and negative strand reads are shown independently. Coverage is more equal and better represented by both strands for the longer sequencing reads.
Figure 4.
Figure 4.
Correlation of probe density and sequencing coverage. Each genomic region was represented on the array with a variable number of capture probes throughout the region. The sequencing coverage per base (blue line) linearly correlates with probe density (red line).
Figure 5.
Figure 5.
Distribution of non-reference allele reads. The percentage of non-reference allele reads was calculated for every heterozygous and homozygous non-reference allele position in the targeted region (n = 1197) and is represented in bins of 5%. For heterozygous calls, the distribution is skewed towards reference allele reads.
Figure 6.
Figure 6.
Sequencing coverage and percentage of non-reference allele distribution for validated and non-validated SNPs. All polymorphic and non-reference positions that were identified by the SNP detection pipeline are plotted as a function of total base coverage versus non-reference read frequency. Validated SNPs (either by their presence in dbSNP or by resequencing) are indicated in blue, non-validated SNPs are shown in red and positions for which no working validation assay could be designed in green. False-positive SNPs tend to have a lower percentage of non-reference allele reads and/or low overall coverage.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Zheng W, Long J, Gao YT, Li C, Zheng Y, Xiang YB, Wen W, Levy S, Deming SL, Haines JL, et al. Genome-wide association study identifies a new breast cancer susceptibility locus at 6q25.1. Nat. Genet. 2009;41:324–328. - PMC - PubMed
    1. Song H, Ramus SJ, Tyrer J, Bolton KL, Gentry-Maharaj A, Wozniak E, Anton-Culver H, Chang-Claude J, Cramer DW, DiCioccio R, et al. A genome-wide association study identifies a new ovarian cancer susceptibility locus on 9p22.2. Nat. Genet. 2009;41:996–1000. - PMC - PubMed
    1. Papaemmanuil E, Hosking FJ, Vijayakrishnan J, Price A, Olver B, Sheridan E, Kinsey SE, Lightfoot T, Roman E, Irving JA, et al. Loci on 7p12.2, 10q21.2 and 14q11.2 are associated with risk of childhood acute lymphoblastic leukemia. Nat. Genet. 2009;41:1006–1010. - PMC - PubMed
    1. Kathiresan S, Voight BF, Purcell S, Musunuru K, Ardissino D, Mannucci PM, Anand S, Engert JC, Samani NJ, Schunkert H, et al. Genome-wide association of early-onset myocardial infarction with single nucleotide polymorphisms and copy number variants. Nat. Genet. 2009;41:334–341. - PMC - PubMed
    1. Ahmed S, Thomas G, Ghoussaini M, Healey CS, Humphreys MK, Platte R, Morrison J, Maranian M, Pooley KA, Luben R, et al. Newly discovered breast cancer susceptibility loci on 3p24 and 17q23.2. Nat. Genet. 2009;41:585–590. - PMC - PubMed

Publication types

Associated data