Zulko
commented
4 years ago Hey, thanks for your interest in DNA Weaver! The library is still undertested for circular sequences afaik, but there should be ways to make it work as you intend.
I believe I understood the issue correctly, and I agree with your approach of providing the origin as a "part" (this I almost equivalent to saying "I have the backbone already"). I am not completely sure why this doesn't work in your example, i.e. why the fragment is not recognized. Do you have a minimal script (ideally with short-ish sequences) to reproduce the issue?
Another and perhaps simpler solution uses the cut_location_constraints parameter to forbid cutting inside a sequence of interest:
uncuttable_sequence = 'CGGGTGCTTGGTGAGACAGCGTATTTTGC'
def cut_location_constraint(sequence):
"""Prevent an assembly station from cutting inside the uncuttable_sequence"""
index = sequence.find(uncuttable_sequence)
if index == -1:
return lambda index: True # allow any cut
start, end = index, index + len(uncuttable_sequence)
return lambda index: not (start <= index <= end) # refuse cuts on that segment
assembly_station = dw.DnaAssemblyStation(
name="Gibson Assembly Station",
assembly_method=dw.GibsonAssemblyMethod(
overhang_selector=dw.TmSegmentSelector(min_tm=55, max_tm=70),
min_segment_length=500,
max_segment_length=4000,
cut_location_constraints=[cut_location_constraint] # <=========
),
supplier=[...],
...
)On another note, if you want to check the solutions returned by DnaWeaver, you can use DnaCauldron, which will take your fragments, your enzyme, and will check that the restriction-ligation will produce the expected construct.
import dnacauldron as dc
repository = dc.SequenceRepository(parts={"f1": fragment_1, "part_B":...})
assembly = dc.Type2sRestrictionAssembly(parts_names)
simulation = assembly.simulate(sequence_repository=repository)
assert len(simulation.construct_records) == 1
predicted_construct_record = simulation.construct_records[0]
from geneblocks import sequences_are_circularly_equal
sequences_are_circularly_equal([predicted_construct_record, my_reference_sequence])
aaroncooper
veghp
Hey-- this is an awesome library, and thanks for developing and sharing it.
I've been playing around with it to think about how I might be able to use it, and I've been testing it by creating a blast db with a bunch of preexisting vectors. I'm testing by trying to build a new one with a small insert, and it looks like the strategies are going to work. However, the origin of those vectors seems to be causing what I consider to be suboptimal solutions. All solutions are creating a junction at the origin with two PCR products rather than making one product going across the origin.
Is this a known issue? I've also tried making a Golden Gate-compatible version of that vector and loading it into a
GoldenGatePartsLibrary, but I still just get the suboptimal PCR product solutions. Do you think I might be doing something wrong?