label atoms in output pdb using bfactor column #483
Conversation
0.25 for unique A 0.50 for core 0.75 for unique B
| # TODO: ??? Protein & crystals == 1.0 | ||
| # cofactor = ??? |
There was a problem hiding this comment.
cofactors should be treated the same was as protein
Codecov ReportPatch coverage:
Additional details and impacted files@@ Coverage Diff @@
## main #483 +/- ##
==========================================
- Coverage 91.99% 91.63% -0.37%
==========================================
Files 106 110 +4
Lines 6324 6542 +218
==========================================
+ Hits 5818 5995 +177
- Misses 506 547 +41
☔ View full report in Codecov by Sentry. |
| @@ -527,10 +527,29 @@ def run(self, *, dry=False, verbose=True, | |||
| ) | |||
|
|
|||
| # b. Write out a PDB containing the subsampled hybrid state | |||
| A_atoms = {hybrid_factory.old_to_hybrid_atom_map[i] | |||
There was a problem hiding this comment.
@IAlibay Is it worth adding this to HTF? Or is it already there and I missed it
There was a problem hiding this comment.
I think everything you need should already be there under:
openfe/openfe/protocols/openmm_rfe/_rfe_utils/relative.py
Lines 439 to 442 in 987009e
So you'd be looking for _atom_classes['unique_new_atoms'] and _atom_classes['unique_old_atoms'] and _atom_classes['core_atoms']
| A_unique = list(hybrid_factory._atom_classes['unique_old_atoms']) | ||
| AB_core = list(hybrid_factory._atom_classes['core_atoms']) | ||
| B_unique = list(hybrid_factory._atom_classes['unique_new_atoms']) | ||
| protein = [] # TODO |
There was a problem hiding this comment.
This isn't where you should be tracking your protein atoms imho. It's a PDB output so you have residue names, those should be used.
As previously mentioned - tracking crystallographic waters is also not an amazing idea since water exchange is a thing.
There was a problem hiding this comment.
yes but even if water exchange is a thing, it's useful to know the origin of the water molecules, which were from addSolvent and which from the initial protein prep. The point here is to be able to understand how each of the Components that were added are propagated into the topology.
| A_unique = list(A_atoms - B_atoms) | ||
| AB_core = list(A_atoms & B_atoms) | ||
| B_unique = list(B_atoms - A_atoms) | ||
| A_unique = list(hybrid_factory._atom_classes['unique_old_atoms']) |
There was a problem hiding this comment.
I'm going to circle back on a similar comment you made last time re: extra assignments - Why do the assignment? Could you not just do the list call in the np.in1d?
There was a problem hiding this comment.
sure, now it's a hideously long line that isn't easy to read
There was a problem hiding this comment.
You could just... break up the line? formatting is a thing
our output PDB topology made it hard to distinguish between the contributions from different ligands. This (mis)uses the bfactor column to label as:
0.25 for unique A
0.50 for core
0.75 for unique B
Then "selecting" A/B is possible in downstream tools