-
|
Given specific predicate def filter_chains(st: gemmi.Structure) -> None:
for model in st:
for chain_index, chain in enumerate(model):
if should_chain_be_deleted(chain):
(???)del model[chain_index]
breakI suspect code above will not work because of model invalidation. |
Beta Was this translation helpful? Give feedback.
Answered by
wojdyr
Mar 27, 2021
Replies: 1 comment 8 replies
-
|
It's similar as it would be with deleting items from a list. Enumerating and deleting at the same time won't work. for i in range(len(somelist) - 1, -1, -1):
if some_condition(somelist, i):
del somelist[i]or you can first make a list of items that are to be deleted, and then delete them in reversed order: to_be_deleted = [n for n, chain in enumerate(model) if should_chain_be_deleted(chain)]
for index in reversed(to_be_deleted):
del model[index] |
Beta Was this translation helpful? Give feedback.
8 replies
-
|
Have you found such a polymer that check_polymer_type() doesn't work? There is no function to delete a polymer only (it hasn't been needed so far). You can delete whole chain or individual residues. |
Beta Was this translation helpful? Give feedback.
-
|
Test looks something like this: def is_rna_entity(entity):
return entity.polymer_type == gemmi.PolymerType.Rna
def is_rna_heur(chain):
return chain.get_polymer().check_polymer_type() == gemmi.PolymerType.Rna
structure = gemmi.read_structure(path)
total_subchain_count = sum(len(entity.subchains) for entity in structure.entities)
rna_subchain_count = sum(len(entity.subchains) for entity in structure.entities if is_rna_entity(entity))
total_chain_count = len(structure[0])
rna_chain_count = sum(1 for chain in structure[0] if is_rna_heur(chain))
warning_flag = "[WARNING]" if rna_subchain_count != rna_chain_count else ""
print(f'[{structure.name}]: By entity: {rna_subchain_count:>2}/{total_subchain_count:<3} By heur: {rna_chain_count:>2}/{total_chain_count:<3} {warning_flag}') It's almost always works, but there are some unsuccessful cases (warnings grep-ed): Output (warnings grep-ed) |
Beta Was this translation helpful? Give feedback.
-
|
That's interesting. I see two main reasons for this.
Some examples for 2. 2BCZ chain A: U C C C DA G U C C A C C G Similarly 3ERC chains (DC U U DC DC). And the other way around. I'm sure inferring the chain type could be further improved, so if anyone has suggestions I'm interested. |
Beta Was this translation helpful? Give feedback.
-
|
Thanks, theoretically, Rna + DnaRnaHybrid should cover all cases. Any hints how to get parent chain for given subchain? Or corresponding polymer entity for given chain (to avoid recalculation if entities already set up)? |
Beta Was this translation helpful? Give feedback.
-
(yes, it's a late answer) |
Beta Was this translation helpful? Give feedback.
Answer selected by
hlvlad
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
It's similar as it would be with deleting items from a list. Enumerating and deleting at the same time won't work.
But it works when going backward. You can either iterate backward (I'm using here example from SO):
or you can first make a list of items that are to be deleted, and then delete them in reversed order: