CONTRA: copy number analysis for targeted resequencing

Abstract

Motivation: In light of the increasing adoption of targeted resequencing (TR) as a cost-effective strategy to identify disease-causing variants, a robust method for copy number variation (CNV) analysis is needed to maximize the value of this promising technology.

Results: We present a method for CNV detection for TR data, including whole-exome capture data. Our method calls copy number gains and losses for each target region based on normalized depth of coverage. Our key strategies include the use of base-level log-ratios to remove GC-content bias, correction for an imbalanced library size effect on log-ratios, and the estimation of log-ratio variations via binning and interpolation. Our methods are made available via CONTRA (COpy Number Targeted Resequencing Analysis), a software package that takes standard alignment formats (BAM/SAM) and outputs in variant call format (VCF4.0), for easy integration with other next-generation sequencing analysis packages. We assessed our methods using samples from seven different target enrichment assays, and evaluated our results using simulated data and real germline data with known CNV genotypes.

Fig. 1.

CONTRA workflow. Either a matched control sample (left arrow) or a pool of normal samples for creating a baseline control (right arrow) must be present.

Fig. 2.

Characteristics of base-level log-ratios. (A) Log-ratio versus GC-content; (B) log-ratio versus log₂-coverage derived from two normal samples; and (C) effect of imbalanced library-size on log-ratios, for both simulated negative binomial data (left) and real data (right). The data points represent copy number neutrality. Top: library size of case sample is two times that of control; middle: equal size; bottom: case is half of control.

Fig. 3.

Comparison of log-ratio variations between matched control and pooled controls of varying number of samples, plotting log-ratio SD against log₂ coverage. The same case sample has been used throughout. Control sample(s) are subset/superset of others.

Fig. 4.

Variation of DOC in TR. (A) Histogram of exon-level coverages in a single sample; (B) coverage profile along a chromosome (showing first 20 k targeted bp of Chromosome 1); (C) coverage versus GC-content; and (D) coverage versus distance from the first targeted base.

Fig. 5.

Coverage correlation between samples. (A) Log-ratio versus targeted base position along Chromosome 20, derived from pairs of random samples as indicated in the plot titles. E.g. top-left: log-ratios between two EZ Exome v2 samples; bottom-right: an EZ Exome v1 sample matched against a SureSelect v2 sample. See also Supplementary Figure S4. (B) Base-level coverage variance against coverage mean, using six random samples for each platform.

Fig. 6.

Receiver operating characteristics (ROC) curve for the HapMap samples, generated by varying CONTRA's P-value threshold. The middle table shows sensitivities and specificities for each individual sample at a P-value of 0.01.