Bigger bacterial genomes do not invest in their gene expression systems...at least not much

Today I was working on a genome catalog of MAGs I extracted from a few metagenomic samples of wild squirrel poop. After assigning KEGG functions to open reading frames (ORFs), I was playing around with the counts of KEGG categories, looking at which categories positively or negatively correlate. I thought it would be a way to look at genomic trade-offs.

At the beginning I plotted correlations of the number of ORFs per category, but it did not show much as all correlations are positive, mainly driven by genome size.

For example, this is the plot of KEGG level 1 categories of spearman correlations:

So I normalized each count by total number of ORFs of the genome and much more interesting results arose:

As we can see we have a lot of negative correlations this time. A lot of them involve "unknown" genes, which in this case are ORFs that lack a KEGG annotation. This makes perfect sense and it is a result of our annotations. The more genes we annotate, of any category, the less unknown.

Other negative correlations involve genetic information processing and organismal systems. The latter seems to be more about eukaryota rather than prokaryota, so we will talk about it now (although I have to look into which KOs are assigned to this, to understand if they are errors or not).

Genetic information processing (GIP) negatively correlates with Metabolism and Environmental information processing. Let's visualize these two correlations.

Pretty good correlations. So you would automatically think that a bacterium with more ORFs dedicated to metabolism or Environmental information processing, has less Genetic information processing...buttttt, it is not really true.

Let's have a look at this plot:

We can see that in bigger genomes (with a higher number of ORFs), normally we see a dramatic increase in the number of genes associated to metabolism and genes to "sense" the environment, but they do not increase much their genetic machinery.

Now why this happens? That would be nice to explain, first starting to know, are bacteria with a bigger genome getting smaller, or vice versa? and it would be a good paper, I think.

code here.