Graphein

Bugfixes

* Remove hydrogen isotopes as well in `graphein.protein.graphs.deprotonate_structure`. [337](https://github.com/a-r-j/graphein/pull/337)
* Fixes bug in sidechain torsion angle computation for structures containing `PYL`/other non-standard amino acids ([357](https://github.com/a-r-j/graphein/pull/357)). Fixes [#356](https://github.com/a-r-j/graphein/issues/356).
* Replaces RCSB PDB FTP urls with new API. [364](https://github.com/a-r-j/graphein/pull/364)
* In Pandas `1.2.0` and later, The default value of regex for `Series.str.replace()` will change from `True` to `False`. So we need use regular expressions explicitly now, to suppress a FutureWarning. By StevenAZy ([359](https://www.github.com/a-r-j/graphein/pull/359))

* Improves the tensor->PDB writer (`graphein.protein.tensor.io.to_pdb`) by automatically unravelling residue-level b-factor predictions/annotations ([352](https://github.com/a-r-j/pull/352)).

* Adds support for PyG 2.4+ ([350](https://www.github.com/a-r-j/graphein/pull/339))
* Fixes `add_sequence_neighbour_vector` to have a zero vector when no neighbor is feasible. Extend to handle insertion codes ([336](https://github.com/a-r-j/graphein/pull/336)).

* Fixes edge case in FoldComp database download if target directory has same name as database ([339](https://github.com/a-r-j/graphein/pull/339))

* Pins BioPandas version to latest

New Features

- [Feature] - [305](https://github.com/a-r-j/graphein/pull/305) Adds the `add_virtual_beta_carbon_vector` function inspired by [RFdiffusion](https://github.com/RosettaCommons/RFdiffusion/blob/main/rfdiffusion/coords6d.py#L37) and ProteinMPNN.

API Changes

- Chain selections are now specified with either `"all"` or a list of strings (e.g. `["A", "B"]`) rather than a single selection string (e.g. `"AB"`). This is a necessary chain due to MMTF support which can have multicharacter chain identifiers. [307](https://github.com/a-r-j/graphein/pull/307)

Improvements

- [Bugfix] - [305](https://github.com/a-r-j/graphein/pull/305) Fixes `add_k_nn_edges` for the case when some residues were dropped before (e.g. when some alt_locs are removed).
- [Bugfix] - [305](https://github.com/a-r-j/graphein/pull/305) Removes obsolete `remove_insertions` in [`rgroup_df` construction](https://github.com/a-r-j/graphein/blob/649a490505740a266b26976807e7f303c2a32ff0/graphein/protein/graphs.py#L540).
- [Bugfix] - [305](https://github.com/a-r-j/graphein/pull/305) Fixes the construction of geometric features when beta-carbons or side chains are missing in non-glycine residues (for example in `H:CYS:104` in 3SE8).
- [Bugfix] - [305](https://github.com/a-r-j/graphein/pull/305) Fixes data types of geometric feature vectors: `object` -> `float`.
- [Bugfix] - [301](https://github.com/a-r-j/graphein/pull/301) Fixes the conversion of undirected NetworkX graph to directed PyG data.
- [Bugfix] - [334](https://github.com/a-r-j/graphein/pull/334) Fixes the corner case of the NetworkX -> PyG conversion when input graph has no edges.

https://github.com/a-r-j/graphein/pull/334

Bugfixes

- Adds missing `stage` parameter to `graphein.ml.datasets.foldcomp_data.FoldCompDataModule.setup()`. [310](https://github.com/a-r-j/graphein/pull/310)
- Ensures exproting groups of PDB chains with PDBManager selects the first model for multu-model structures. [311](https://github.com/a-r-j/graphein/pull/311)
- Fixes bug with exporting PDBs with only one splitting strategy in PDBManager [311](https://github.com/a-r-j/graphein/pull/311)
- Fixes incorrect jaxtyping syntax for variable size dimensions [312](https://github.com/a-r-j/graphein/pull/312)
- Fixes shape of angle embeddings for `graphein.protein.tesnor.angles.alpha/kappa`. [315](https://github.com/a-r-j/graphein/pull/315)
- Fixes initialisation of `Protein` objects. [317](https://github.com/a-r-j/graphein/issues/317) [#318](https://github.com/a-r-j/graphein/pull/318)
- Fixes incorrect `rad` and `embed` argument logic in `graphein.protein.tensor.angles.dihedrals/sidechain_torsion` [321](https://github.com/a-r-j/graphein/pull/321)
- Fixes incorrect start padding in pNeRF output [321](https://github.com/a-r-j/graphein/pull/321)
- Fixes `pyyaml` breaking installation [328](https://github.com/a-r-j/graphein/pull/328)
- Fixes setting ID for PyG data objects when loading from a path to a `.pdb` file [332](https://github.com/a-r-j/graphein/pull/332)

Other Changes

- Adds transform composition to FoldComp Dataset [312](https://github.com/a-r-j/graphein/pull/312)
- Adds entry point for biopandas dataframes in `graphein.protein.tensor.io.protein_to_pyg`. [310](https://github.com/a-r-j/graphein/pull/310)
- Adds support for `.ent` files to `graphein.protein.graphs.read_pdb_to_dataframe`. [310](https://github.com/a-r-j/graphein/pull/310)
- Obsolete residues with no replacement are now returned by `graphein.protein.utils.get_obsolete_mapping`. [310](https://github.com/a-r-j/graphein/pull/310)
- Adds the ability to store a dictionary of HETATM positions in `Data`/`Protein` objects created in the `graphein.protein.tensor` module. [307](https://github.com/a-r-j/graphein/pull/307)
- Improved handling of non-standard residues in the `graphein.protein.tensor` module. [307](https://github.com/a-r-j/graphein/pull/307)
- Insertions retained by default in the `graphein.protein.tensor` module. I.e. `insertions=True` is now the default behaviour.[307](https://github.com/a-r-j/graphein/pull/307)
- `plot_pyg_data` now also plots some geometric features if present. [305](https://github.com/a-r-j/graphein/pull/305)
- Adds transform composition to FoldComp Dataset [312](https://github.com/a-r-j/graphein/pull/312)
- Improve FoldComp dataloading performance and include B factors (pLDDT) in output. [313](https://github.com/a-r-j/graphein/pull/313) [#315](https://github.com/a-r-j/graphein/pull/315)
- Add new helper functions to PDBManager [322](https://github.com/a-r-j/graphein/pull/322) (amorehead)
- Add non-standard 'CYX' to `RESI_THREE_TO_1`.