FIG 4.

Compendium filtering guided by distributions of sparsity and housekeeping gene expression. (A) Filtering thresholds for sparsity were defined as the 10th and 90th percentiles for PAO1 and PA14, separately determined for samples annotated as belonging to each strain, as indicated by circles showing PAO1 (dark purple), PA14 (light purple), or otherwise annotated or unannotated samples (gray). Samples were excluded (removed) (orange-filled circles) from the compendia if they fell outside the ranges by both PAO1- and PA14-defined criteria; otherwise, they were included (maintained) (unfilled circles). The top and side histograms show profile counts. (B) Filtering thresholds for housekeeping gene expression were defined as the 20th and 98th percentiles using the same samples as the ones used for the sparsity threshold determinations. Strain annotations and inclusion (maintained) or exclusion (removed) are indicated by the circle outlines and fill colors as described above for sparsity filtering. The top and side histograms show profile counts.