Changelog
Additions
- Added interfaces to other packages in the bioinformatics ecosystem via the new
interface
subpackage. (#762)interface.rdkit
allows conversion betweenAtomArray
andMol
objects from the RDKit package. (#736, #741, #742)interface.pymol
allows conversion betweenAtomArray
and PyMOL objects, supporting NumPy-style indices as atom selections. (#749)interface.openmm
allows conversion betweenAtomArray
and OpenMMToplology
,System
andState
objects. (#750)
- Added
database.alphafold
subpackage for accessing structures from AlphaFold DB. (#492)- The IDs of computational models obtained via
database.rcsb.search()
can be passed directly todatabase.alphafold.fetch()
.
- The IDs of computational models obtained via
- Added new metrics for measuring similarity of two structures.
- Added functions related to aromatic rings. (#751)
structure.find_aromatic_rings()
finds atom indices of aromatic rings in a structure.structure.find_stacking_interactions()
finds pi-stacking interactions between aromatic rings.
- Added
structure.BondType.AROMATIC
bond type, for bonds where the kekulized form is unknown. (#738) structure.info.residue()
andstructure.io.pdbx.get_component()
now support chemical components without valid coordinates. (#730)label_entity_id
annotation can now be written to PDBx files viastructure.io.pdbx.set_structure()
. (#732)structure.info.vdw_radius_single()
now supports more elements. (#745)- Improved performance of
structure.spread_residue_wise()
andstructure.spread_chain_wise()
. (#770)
Fixes
- Now only finite coordinates can be given to
structure.CellList()
(#733). - Fixed
structure.array()
truncating string annotations. (#755) structure.io.pdbx.get_structure()
withinclude_bonds=True
is now much faster and more memory efficient for very large structures. (#765)- Fixed serialization (and string representation) of
structure.io.pdbx.CIFBlock
objects raising an exception. (#767) - Now
structure.io.mol.set_structure()
checks if the input coordinates fit into the fixed number of columns in MOL/SDF files. (#772) - The height of letters in
sequence.graphics.plot_feature_map.plot_sequence_logo()
is now more accurate. (#769)