I’m kicking around an idea that involves a genetic screen with a ~30k variant library. It would be insanely beneficial if I could assay by targeted sequencing (like in a crispr screen), but instead of retrieving the variants directly, retrieve a barcode.

If I were to do that, I would need to build my screening construct (a ~25kb plasmid), which would contain the variant library and barcode. The barcode would be separated from the variant library by ~8 kb, and I’d want 1 barcode per variant. I’d synthesize the sequence encoding these variants/barcodes and clone them into the vector. How would you generate the genetic link between variant and barcode (I.e., the 1:1 relationship between them, where each plasmid contains 1 variant and its unique barcode, separated by ~8 kb of other sequence)?

I’ve kicked around golden gate assembly where the variant and barcode are encoded on the same oligo within an oligo library. Presumably because intramolecular rxns are faster, this might work. Alternatively, I can imagine some sort of emulsified golden gate or Gibson-like reaction where each droplet contains (on average) 0.3 oligos and 0.3 copies of plasmid backbone with an excess of cloning reagents (meaning few droplets have both, and almost none have >1 oligo and plasmid backbone).

What are y’all’s ideas? Is anyone aware of a method that already exists and is well-established?