NAHRly normalizes read depth in a given genomic region across multiple samples and then uses the troughs of the histogram of normalized read depths to threshold them, thereby clustering the samples into copy-number classes. (This approach is borrowed from the field of image segmentation .)

Using the labels assigned to samples by the histogram-based method outlined above, Naive Bayes was used to compute

• confidence in the histogram-based inference of copy numbers
• re-predict copy numbers

This dashboard visualizes the Naive-Bayes confidence of inferred copy numbers and compares the histogram-based and Naive-Bayes-based predicted copy numbers.