Figure 6

Simulations. Effect of sample contamination and read depth estimation errors on a simulated cancer genome. ϕ is a scaling factor for the variance; for example ϕ = 400 means that the noise model has a standard deviation 20 times r _j for interval I _j . We show the average percent of interval edges (left) and reference and variant edges (middle) correctly estimated over 10 trials. (Right) At ρ = 0, as ϕ increases most of the errors result from variant edges moving from the correct multiplicity of 1 (heterozygous deletion) to a multiplicity of 2 (homozygous deletion).