Again. It's the molecule that is patented, not the process.
I'm going to say it once more. It's the molecule that is patented. Not the information. Not the process. The MOLECULE.
"If the gene information previously exists in nature, ALL equivalent forms (DNA, mRNA, cDNA, even tRNA) should be excluded from patentability, because they are beyond obvious once you know the nucleotide sequence, and they are equivalent (in terms of information theory). "
Fine, that's what you believe. However, that is not what the law is. Molecule patents say nothing about information, "equivalent" forms, and have a different definition of 'obvious' than you do. There's a very good reason for them to be defined the way they are, too, because you have to be careful when dealing with legalities. The law has to be (and often fails at being) explicit with these concepts that can be gray, because the way you're treating them is very ill-defined.
Your wooden machine is different from cDNA. cDNA is an obvious 1:1 mapping of something that already exists: mRNA. It's obvious because that's exactly what retroviruses do. (Ignoring here the negative vs positive sense distinction in viral RNA because it's irrelevant once the RNA is converted to double-stranded DNA.)
If the gene information previously exists in nature, ALL equivalent forms (DNA, mRNA, cDNA, even tRNA) should be excluded from patentability, because they are beyond obvious once you know the nucleotide sequence, and they are equivalent (in terms of information theory). The equivalence (ignoring introns) of all forms of DNA and RNA is universally known (barring strange cases where DNA/RNA structure might interfere with transcription).
Novel synthesis methods might be patentable. Novel enzymes that can translate between the forms, also might be patentable. But that's not this case. Often when something obvious or preexisting is patented, a retort is that while the thing may not be patentable, the process to create it might be. I was trying to head off that argument that Myriad might have something patentable here even if the cDNA itself is not patentable.