While the Val158Met allele (rs4680) has been well-studied and appears to affect brain function, there is evidence that COMT activity levels are largely controlled by a less-obvious mechanism. [PMID 15537663] studied rs6269, rs4633, rs4818, and rs4680, and grouped COMT function into three haploblocks based on pain response, which covered about 96% of the people tested:
-people in the GCGG block showed the lowest pain response (LPS), attributed to high COMT function; so the lowest pain sensitivity group are individuals carrying two copies of this haplotype, designated LPS/LPS
-ATCA people had intermediate (average) pain response (APS); individuals in the average category have either APS/APS, HPS/LPS, or LPS/APS diplotypes
-ACCG people had the highest pain response (HPS), attributed to low COMT function; individuals in this category are APS/HPS or HPS/HPS
Note that two of these SNPs are "silent" or synonymous variations, one is a promotor-region variation, and that the Val158Met variation, rs4680(G) appears in both the highest-function and lowest-function groups. This suggests an activity profile that is not affected by amino acid sequence, but perhaps rather by some quality of the RNA transcript. The study found similar quantities of RNA transcribed for all three haploblocks, but different levels of COMT function, suggesting that the mechanism may lie in translational efficiency.
[PMID 18698234] 4 SNPs affecting COMT activity were used to define low and high activity haplotypes. These SNPs are: rs6269, rs4633, rs4818, and rs4680. The mean levodopa dose increased in correspondence with the haplotype activity (low < medium < high). Doses prescribed for G-C-G-G (high activity) haplotype carriers (mean 604.2+/-261.9 mg) were significantly higher than those for the noncarriers (mean 512.2+/-133.5 mg, p<0.05).