Table 1 Advantages and disadvantages of using relative read depths vs. VAF for determining CNV

Read depth

1. Uses any locus with reasonable coverage, providing higher resolution.

2. Dependent on proper mapping, not on variant calls.

3. Relatively robust to small contamination events from another human sample.

1. Highly susceptible to technical bias both from batch to batch and from sample to sample due to variable sensitivity of the various regions to moderate or even slight changes in hybridization conditions and other factors (such as exome enrichment kit versions). When used by itself, this approach can yield a large number of false positives and negatives if not sufficiently validated.

2. Unable to identify complex copy number events such as CN-LOH or hyperdiploidy.

Variant allele frequency (VAF)

1. Uses heterozygote positions to discern bands that indicate larger copy number changes.

2. More resistant to technical bias from batch to batch and from sample to sample.

1. Can confuse copy number events with contamination events.

2. Cannot discern concurrent gains in both alleles from normal state.

3. Misinterprets CN-LOH events as copy number loss events.

4. Requires that variant calling and heterozygous states be established. Resolution depends on the number and distribution of variants called.

Read depth combined with VAF

1. One method fills in the gaps in information from the other method.

2. Properly assesses CN-LOH events and concurrent gains in both alleles as well as copy number state when small levels of contamination occur.

1. Inherently different resolution levels complicate creation of individual segments from both sources of information.