BreakDancer: an algorithm for high-resolution mapping of genomic structural variation

Abstract

Detection and characterization of genomic structural variation are important for understanding the landscape of genetic variation in human populations and in complex diseases such as cancer. Recent studies demonstrate the feasibility of detecting structural variation using next-generation, short-insert, paired-end sequencing reads. However, the utility of these reads is not entirely clear, nor are the analysis methods with which accurate detection can be achieved. The algorithm BreakDancer predicts a wide variety of structural variants including insertion-deletions (indels), inversions and translocations. We examined BreakDancer's performance in simulation, in comparison with other methods and in analyses of a sample from an individual with acute myeloid leukemia and of samples from the 1,000 Genomes trio individuals. BreakDancer sensitively and accurately detected indels ranging from 10 base pairs to 1 megabase pair that are difficult to detect via a single conventional approach.

Figure 1

Overview of BreakDancer algorithm. (a) The workflow. (b) Five types of anomalous read pairs recognized by BreakDancerMax. A pair of arrows represents the location and the orientation of a read pair. A dotted line represents a chromosome in the subject genome. A solid line represents a chromosome in the reference genome.

Figure 1

Overview of BreakDancer algorithm. (a) The workflow. (b) Five types of anomalous read pairs recognized by BreakDancerMax. A pair of arrows represents the location and the orientation of a read pair. A dotted line represents a chromosome in the subject genome. A solid line represents a chromosome in the reference genome.

Figure 2

Performance of BreakDancer in simulation. TPR and FPR of BreakDancerMax (BDMax) at the confidence threshold of Q ≥ 30 are shown. TPR analytic refers to the percent of variants that can hypothetically be detected by BDMax under an analytic model (Online Methods). TPR detectable is the percent of variants hose flanking regions (300 bp both to the left and to the right) contain 2 or more confidently mapped ARPs in the MAQ alignment. The performance of BreakDancerMini (BDMini) is characterized by its TPR and FPR. The combined performance (BD all) is obtained by merging the results of these to programs.

Figure 3

Size distribution of deletions detected in an AML genome. 3170 deletions were detected from the sequence data by BreakDancerMax ranging from 58 bp to 959,498 bp. To signature peaks at 300 bp and at 6,000 bp correspond respectively to the AluY and the L1Hs retro-transposon. In comparison, only 116 inherited CNVs were detected using Affymetrix 6.0 array on this sample.

Figure 4

Accuracy of predicted variant sizes. Plotted are variant sizes predicted by BreakDancer and by local assembly versus true sizes determined from the PCR resequencing (x axis). Positive sizes represent deletions and negatives represent insertions.