Interactive version of Figure 1:

Chromosomes and ERs distribution by GC content and replicon size

We compiled all replicons from complete bacterial genomes and classified them into two categories:

• chromosomes – the largest replicon in each genome
• ERs – all remaining replicons

The rationale behind this classification is described in the Methods, Replicon analysis section.

For every taxonomic group at the levels of phylum, class, order, family, and genus that contained ≥100 ERs in our dataset, we generated two-dimensional scatter plots and kernel density (KDE) plots represented as violin plots using GC content (%) and replicon size (bp).

For each eligible taxonomic group, two versions of every plot were produced:

• one with size shown on a linear scale
• one with size shown on a log₁₀ scale

KDE violin plots were scaled identically within each taxon to enable direct comparison of the shape of chromosome and ER distributions, rather than their absolute counts. KDE is a non-parametric technique that estimates smooth probability-density curves without assuming normality or unimodality (Kitchenham, 2015; Węglarczyk, 2018).

References

• Kitchenham B. 2015. Robust statistical methods: why, what and how. Proceedings of the 19th International Conference on Evaluation and Assessment in Software Engineering (EASE ’15), 1–6.
• Węglarczyk S. 2018. Kernel density estimation and its application. ITM Web Conf. 23:00037.

Phyla with ≥100 ERs

Classes with ≥100 ERs

Orders with ≥100 ERs

Families with ≥100 ERs

Genera with ≥100 ERs