Well, yeah, if what he's talking about are nucleotide bases, it takes three of them to make a codon. So if you have a sequence ABCABC, then ABC might be a codon (which codes for an amino acid, like a "bead" in a necklace that makes up a protein), BCA might be a codon, and CAB might be a codon. Then at the same time you could go backwards. So really there's six potential different codons that a single basepair might be involved in. But that's probably not what this paper is talking about.

You're correct in that this is not what the paper is talking about. However, there is a third layer (at least) beyond the gene-coding, most-studied area, and the article submitted; epigenetics studies how gene expression is modified by histone modification and DNA methylation, basically different ways of switching gene expression on/off or changing the amount of protein made from a certain gene.