Marketing Experiments did a series of tests for clients to show the effect of copy length on Website conversion rates.
For the 20 clusters with the best score ( = total copy length * copy pair amount), we performed a motif search with at least 60% sequence identity on both strands and in both directions.
Their inherent features are copy length (or right and left copy length for imperfect IRs), spacer length, degree of perfection, percent A+T composition, dinucleotide repeat content, percent match, percent insertion deletion, and IRF score.
The longest perfect IR found in the S. cerevisiae genome had copy length 30, while the average number of perfect IRs with copy length larger than 15 was less than one both in the R and in the RA genomes.
For repeat numbers n = 1 or 2, the IR count is weakly positively correlated with copy length.
Using the IRFinder algorithm we detect 100,514 IRs having copy length greater than 6 bp and spacer length less than 77 bp.
At every copy length the perfect IR number is statistically significantly higher in the S. cerevisiae genome than in either the R or the RA genomes.
So a portion of the observed slightly higher frequency of perfection in the non-overlapping set may be a consequence of its shorter average copy length.
In contrast, the longest overlapping IR within any of the 100 randomized genomes (either R or RA) has copy length of 92 bp.
The mean copy length for the former is μ S. cerevisiae ≈ 14.7 bp, while for the RA randomized genomes it is μRA ≈ 14.0 bp.
