evolution - Is there variation of AT/CG ratio along species?

That ratio is essentially, as WYSIWYG pointed out, called GC-content. In actuality, GC-content is reported as $(G+C)/(A+C+G+T)$, converted to percent; i.e., what percent of the genome is G or C.

There is vast variation in GC-content, both amongst species and within a given species' own genome. For example, in humans the first intron and exon are generally more GC-rich than following introns/exons.¹ Genes themselves are often found in higher GC areas,^2,11 in particular CpG islands are found near a large number of (mammalian) promoters.³

Across species, there can be a big difference. Yeast and Arabidopsis are both around 35%^4,5 whereas Plasmodium falciparum is around 24%;⁶Carsonella are even lower, at around 16.5%.¹⁴ On the other hand, the plankton Emiliania huxleyi is around 65%⁷. We can use these differences to study genomic history. Bacteria often have genes from all over the place thanks to horizontal gene transfer, and GC-content can be used to differentiate between their own genes and those from horizontal gene transfer;⁸ a good example is the CRISPR-Cas system,⁹ even in a virus!¹⁰

Here's a list of a few things genomic GC-content is correlated with:¹⁵

• genome size


• whether the bacterium is free-living or not


• the environment


• aerobiosis


• nitrogen utilization

In the lab, high GC-content often means a harder region to work with, as the presence of three instead of two bonds (between A and T) requires more energy to break;¹² anything involving primers can be made more annoying, including (especially, to some) sequencing. There is a theory that high GC-content would be an adaptation to high temperatures, to avoid DNA damage, but that is controversial.^13,16,17,18