Closed dave-doty closed 3 years ago
Paranemic crossovers are possible in cadnano thus bug must be internal to conversion algorithm.
I added tests_inputs/cadnano_v2_import/test_paranemic_crossover.json
which is valid cadnano file with 2 paranemic crossovers (one on staples, one on scaffold).
tests_inputs/cadnano_v2_import/test_paranemic_crossover.json
in scadnano worksAfter import then export:
Bug fixed:
Closed in #189.
scadnano allows "paranemic" crossovers (https://doi.org/10.1021/ja038381e, https://doi.org/10.1021/acs.chemrev.8b00207), i.e., crossovers that connect two domains in the same direction:
Exporting to cadnano (Export --> cadnano v2 from menu) and then reimporting that same
.json
file (File --> Import cadnano v2 from menu) gives an error:The first domain mentioned in the error exists, and is the 3'-most domain of the scaffold. But the second domain (helix=2, start=15, end=16) does not exist in the original design.
The scadnano file with this design is
Note that because the above strand is a scaffold, and we want the strand to be forward on both helices, we must have both helices be even due to cadnano limitations on directions scaffold/staple strands can go on helices based on their parity.
However, the same error happens if the strand is a staple, and is reversed on both helices: