> It's very hard to Figure out what a random piece of DNA will do just from the sequence.
While it's certainly true that it's hard to determine the function of a DNA sequence from scratch, it's considerably easier to compare that sequence (or the sequence of the translated polypeptide) to other homologous sequences to see if it matches something dangerous.
I previously worked in a lab that studied Bacillus anthracis, and we had a bit of trouble getting a major gene synthesis company [1] to produce a plasmid with a variant of atxA [2], and atxA isn't even a toxin, it's just a transcriptional regulator. We presumed that they just BLASTed [3] the sequence we gave them and threw up a red flag when it matched anthracis. So this sort of sequence-checking already occurs.
While it's certainly true that it's hard to determine the function of a DNA sequence from scratch, it's considerably easier to compare that sequence (or the sequence of the translated polypeptide) to other homologous sequences to see if it matches something dangerous.
I previously worked in a lab that studied Bacillus anthracis, and we had a bit of trouble getting a major gene synthesis company [1] to produce a plasmid with a variant of atxA [2], and atxA isn't even a toxin, it's just a transcriptional regulator. We presumed that they just BLASTed [3] the sequence we gave them and threw up a red flag when it matched anthracis. So this sort of sequence-checking already occurs.
[1] https://www.dna20.com/
[2] http://www.ncbi.nlm.nih.gov/pubmed/8577251
[3] http://blast.ncbi.nlm.nih.gov/