Fig. 3

Comparison of the different methods. The y-axis shows on a logarithmic scale the total number of false positive seeds that are scaled to be estimates of the total number of false positive seeds per base in the genome if two complete human-sized genomes were compared (\(\mu\) means \(10^{-6}\)). The x-axis is the percentage of coding exons that are supported by seeds. Each data point represents a run of the respective method with a certain weight k (point labels), ranging from 12 (top right points) to 24 (bottom left). Note that data points at 0 \(\widehat{\text {FP}}\) are slightly shifted for better visibility. The filtering to consider at most 10 seed candidates per k-mer was relaxed to 100 for all weight 12 runs