So here's my shot in the dark; you found reasons to draw a new distinction between C-linked and N-linked, one that is previously unrecognized by the mainstream, which describes their relationship as follows: C-linked being upstream from N-linked and more genetically conserved...?It certainly attributes it a meaning, but doesn't of course change the observations that 'methylation' in general acts to suppress a gene. The main point that I try to make is there is a fundamental difference between C‐linked and N‐linked methylation. While N‐linked methylation could be considered a post‐translational modification, C‐linked 'methylation' is non‐labile and had actually existed the moment the DNA strand was synthesized. The reason I say this must be the case is: The enzyme assumed responsible for this has dismal kinetic rates, impossible reaction mechanisms, and appears to have never actually been demonstrated as such. Any radioactive methyl groups found transferred—at a rate of about one per hour [sic]—through experiment could very well have been the result of a low‐energy nitrogen methylation, a completely realistic event. The enzymatic formation of a carbon–carbon bond usually requires a transition element in the enzyme's catalytic domain, something cytosine‐5‐methyltransferase does not have; some biochemists appear to think they can just wave a magic wand to explain these irregularities, but they don't even have to. The methyl group has its origin before the DNA nucleotide cytosine had even been synthesized, formed the moment glutamate was isomerized to β-methylasparatate. Regular cytosine is synthesized from aspartate, and methylcytosine is synthesized by β-methylasparatate. All than needs to be done to explain the C‐linked methylation is to take the consensus scheme for cytosine biosynthesis and instead run β-methylasparatate through that very scheme. The result of this is, of course, 5-methylcytosine; I'm convinced that this is how it's formed. This provides a more grounded and reasonable explanation than the impossible-sounding feats and low kinetic rates of the putative enzyme cytosine-5-methyltransferase.
Accepting this leads to the idea that the ratio of β-methylasparatate/asparate alone is fundamentally responsible, and drives so-called C‐linked 'methylation.' And since β-methylasparatate is isomerized from glutamate, this is also dependent on the β-methylasparatate/glutamate ratio. You could even then algebraically substitute one of these in the other to deduce the suspicion that the glutamate/asparate ratio can be thought responsible.
Confirmation of this idea comes in the observation that genes which control enzymes involved in glutamate and aspartate metabolism are CpG islands—areas of very high C-linked DNA methylation. Accepting all of this leads to the natural conclusion that the glutamate/asparate ratio itself regulates the very genes which control its metabolism through the β-methylasparatate/asparate ratio, the 5-methylcytosine/cytosine ratio, the DNA methylation itself, and finally: the suppression of genes in a manner which brings the glutamate/asparate ratio back to equilibrium.