You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Maybe add a version attribute in DSSPConfig? or automatically detect it from running subprocess dssp --version which returns mkdssp version 4.0.4 for me.
The text was updated successfully, but these errors were encountered:
Hi @biochunan, appreciate you raising this issue. I'm currently loathe to support DSSP 4 as it requires CRYST1 records for PDB files, which while sensible, will break compatibility for a lot of people who use Graphein for predicted structures. As a result, I'd suggest the conda distribution of DSSP provided by the Sali lab. I realise this is not an ideal strategy in the longterm but am not able to prioritize this at the moment, if you'd like to make a PR, I'm happy to assist - as you might be able to tell from this PR I've already tried to setup a more universal solution. Automatically detecting the version seems like the best approach.
Hello!
https://github.com/a-r-j/graphein/blob/3d7af1fd07dc16707068b2d63a2f3668fc02c632/graphein/protein/features/nodes/dssp.py#L110C71-L110C71
Here, the
dssp
version is not specified, which leads to Bio.PDB.DSSP uses the default version3.9.9
for dssp. For people using dssp version>=4.0.0
this leads to the output format by default asmmcif
.I included
Bio.PDB.DSSP
line 135 for referenceBecause I'm using dssp version 4.0.4 and without specifying the version number, it leads to an empty
dssp_df
, and adding an arg ofdssp_version
works.Maybe add a
version
attribute inDSSPConfig
? or automatically detect it from running subprocessdssp --version
which returnsmkdssp version 4.0.4
for me.The text was updated successfully, but these errors were encountered: