Copy Number predictions across the genomes were correlated with the read counts and their fold difference in coverage against a median (~1 copy) coverage. Plots below demonstrate the effect of correction of bias in read counts in determining the number of natural copies being sequenced. The HMM we adapted for this process demonstrates congruence with the expected results based on the changes in corrected coverage. Determining accuracy of the model would require known characterized true positive samples, however the LTEE clones haven't been characterized for CNVs. We predict testing additional samples and characterizing more CNV data points will help evaluate and improve the model further.
Figure 10: Plots above demonstrate the copy number calls (right y-axis) with respect to the read counts (left y-axis) observed per sample. The red points correspond to the copy number predictions made using HMM at specific genomic locations.