Talk:Synthetic Biology:Vectors/Modular construction scheme
From OpenWetWareJump to navigationJump to search
- Austin 13:15, 10 March 2006 (EST): I think it kind of depends on how you view the function of the barcode. If it's something like an inline comment, then putting the barcode with the part itself would be fine and it would be inserted and removed along with the rest of the part. If it's more like system level documentation where there's one location where the documentation should go, then you have the same problem as in any other system like software. But it doesn't make it a bad idea to have documentation even if there's no way to force people to update the documentation when they change the code. That's why people are tending to move comments with the code and then generating system level documentation automatically. I was thinking that it would be impossible to do this with biology but here's a random idea: you flank these barcodes with a normal symmetric restriction enzyme. Thus when you cut with this enzyme you cut all these barcodes (and perhaps other stuff, doesn't really matter). Everything leaves the same end, you ligate the whole mess together with some known linkers also, and optimize for many barcode ligations before the linkers do. Lots of different things will form which you then PCR and then sequence. You can calculate the chances that you'll get all the different barcodes within a sequencing run. For example, if there are 5 different barcodes located throughout the plasmid each of ~50 bp, then it seems reasonable to get these 250 bp somewhere in a 900 bp read.
- RS 13:24, 10 March 2006 (EST): So to date, we've been talking about barcodes primarily as something in a single location rather than peppered throughout the plasmid like inline comments. This is likely because the initial motivation for barcodes was to be able to identify the vector during sequencing with the verification primers. Now I could barcode each part of the plasmid, but I am not sure how much this buys you practically speaking since vector identification becomes a fair bit of work (under your proposed scheme). If we do go with the scheme for documentation at the system level in a single location, I am not sure that incorrect documentation is in fact better than no documentation in this case.