Skip to content

Latest commit

 

History

History
342 lines (238 loc) · 24.5 KB

CHANGELOG.md

File metadata and controls

342 lines (238 loc) · 24.5 KB
Raw

1.7.7 - UNRELEASED

Bugfixes

  • Add support for DSSP >4. Backwards compatibility is still supported. #355. Fixes #353.
  • Fixes bug where RSA features are missing from nodes with insertion codes. #355. Fixes #354.
  • Fix bug where the deprotonate argument is not wired up to graphein.protein.graphs.construct_graphs. #375

Misc

  • bumped logging level down from INFO to DEBUG at several places to reduced output length #391
  • exposed fill_value and bfactor option to protein_to_pyg function. #385 and #388
  • Updated Foldcomp datasets with improved setup function and updated database choices such as ESMAtlas. #382
  • Resolve issue with notebook version and pluggy in Dockerfile. #372
  • Remove typing_extension as dependency since we now primarily support Python >=3.8 and Literal is included in typing there.

1.7.6

Bugfixes

  • Fixes bug in pdb_manager for clustering sequences via mmseqs #377
  • Remove hydrogen isotopes as well in graphein.protein.graphs.deprotonate_structure. #337
  • Fixes bug in sidechain torsion angle computation for structures containing PYL/other non-standard amino acids (#357). Fixes #356.
  • Replaces RCSB PDB FTP urls with new API. #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)

1.7.5 - 27/10/2024

  • Improves the tensor->PDB writer (graphein.protein.tensor.io.to_pdb) by automatically unravelling residue-level b-factor predictions/annotations (#352).

1.7.4 - 26/10/2023

  • Adds support for PyG 2.4+ (#350)
  • Fixes add_sequence_neighbour_vector to have a zero vector when no neighbor is feasible. Extend to handle insertion codes (#336).

1.7.3 - 30/08/2023

  • Fixes edge case in FoldComp database download if target directory has same name as database (#339)

1.7.2 - 28/08/2023

  • Pins BioPandas version to latest

1.7.1 - 26/07/2023

New Features

  • [Feature] - #305 Adds the add_virtual_beta_carbon_vector function inspired by RFdiffusion 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

Improvements

  • [Bugfix] - #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 Removes obsolete remove_insertions in rgroup_df construction.
  • [Bugfix] - #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 Fixes data types of geometric feature vectors: object -> float.
  • [Bugfix] - #301 Fixes the conversion of undirected NetworkX graph to directed PyG data.
  • [Bugfix] - #334 Fixes the corner case of the NetworkX -> PyG conversion when input graph has no edges.

#334

Bugfixes

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

Other Changes

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

1.7.0 - 10 /04/2023

New Features

  • [PDBManager] - #272 Adds a utility for creating custom dataset splits from the PDB.
  • [FoldComp Dataset] - #284 - Create ML datasets from FoldComp databases.
  • [ESM] - #284 - Wrapper for ESMFold batch folding & embedding.
  • [Downloads] MMTF downloading now supported in download utilities. #272

API Changes

  • The pdb_path argument to many functions (e.g. graphein.protein.graphs.construct_graph) has been renamed to path as this can now accept MMTF files in addition to PDB files.
  • Protein tensors have coordinates renamed from Protein.x to Protein.coords. #272

Other changes

  • Tensor types are now defined using jaxtyping, removing the torchtyping dependency #272
  • Drops explicit Python 3.7 support. Colab now runs on 3.8+. #272
  • Dockerfile now builds from pytorch/pytorch:1.13.0-cuda11.6-cudnn8-runtime (replaces pytorch/pytorch:1.9.1-cuda11.1-cudnn8-runtime) #272
  • Missing os import fixed in [#297(#297). Fixes #296

1.6.0 - 18/03/2023

New Features

  • [Metrics] - #245 Adds a selection of structural metrics relevant to protein structures.
  • [Tensor Operations] - #244 Adds suite of utilities for working directly with tensor-based representations of proteins (graphein.protein.tensor).
  • [Tensor Operations] - #244 Adds suite of utilities for working with ESMfold (graphein.protein.folding_utils).

Improvements

  • [Feature] = #277 Adds support for pathlib paths for protein graph creation. #269
  • [Logging] - #221 Adds global control of logging with graphein.verbose(enabled=False).
  • [Logging] - #242 Adds control of protein graph construction logging. Resolves #238

Protein

  • [Bugfix] - [#222]#222 Fixes entrypoint for user-defined df_processing_funcs (#216)

  • [Feature] = #263 Adds control of Alt Loc selection strategy. N.b. Default ProteinGraphConfig changed to include insertions by default (insertions=True) and alt_locs="max_occupancy".

  • [Feature] - #264 Adds entrypoint to graphein.protein.graphs.construct_graph for passing in a BioPandas dataframe directly.

  • [Feature] - #229 Adds support for filtering KNN edges based on self-loops and chain membership. Contribution by @anton-bushuiev.

  • [Feature] - #234 Adds support for aggregating node features over residues (graphein.protein.features.sequence.utils.aggregate_feature_over_residues).

  • [Bugfix] - #234 fixes use of nullcontext in silent graph construction.

  • [Bugfix] - #234 Fixes division by zero errors for edge colouring in visualisation.

  • [Bugfix] - #254 Fix peptide bond addition for all atom graphs.

  • [Bugfix] - #223 Fix handling of insertions in protein graphs. Insertions are now given IDs like: A:SER:12:A. Contribution by @manonreau.

  • [Bugfix] - #226 Catches failed AF2 structure downloads #225

  • [Bugfix] - #229 Fixes bug in KNN edge computation. Contribution by @anton-bushuiev.

  • [Bugfix] - #220 Fixes edge metadata conversion to PyG. Contribution by @manonreau.

  • [Bugfix] - #220 Fixes centroid atom grouping & avoids unnecessary edge computation where none are found. Contribution by @manonreau.

  • [Bugfix] - #268 Fixes 'sequence' metadata feature for atomistic graphs, removing duplicate residues. Contribution by @kamurani.

ML

  • [Bugfix] - #234 - Fixes bugs and improves conversion.convert_nx_to_pyg and visualisation.plot_pyg_data. Removes distance matrix (dist_mat) from defualt set of features converted to tensor.

Utils

  • [Improvement] - #234 - Adds parse_aggregation_type to retrieve aggregation functions.

RNA

  • [Bugfix] - #281 - Bugfix for nx->PyG conversion for graphs containing edges without "kind" attributes. Contribution by @rg314.

Constants

  • [Improvement] - #234 - Adds 1 to 3 mappings to graphein.protein.resi_atoms.

Documentation

  • [Tensor Module] - #244 Documents new graphein.protein.tensor module.
  • [CI] - #244 Updates to intersphinx maps

Package

  • [CI] - #244 CI now runs for python 3.8, 3.9 and torch 1.12.0 and 1.13.0
  • [CI] - #244 Separate builds for core library and library with DL dependencies.
  • [Licence] - #244 Bump to 2023

1.5.2 - 19/9/2022

Protein

  • [Bugfix] - #206 Fixes KeyError when using graphein.protein.edges.distance.node_coords
  • [Bugfix] - Includes missing data files in MANIFEST.in #205

GRN

  • [Bugfix] - #208 - Resolves SSL issues with RegNetwork.

ML

  • [Feature] - #208 support for loading local pdb files by ProteinGraphDataset and InMemoryProteinGraphDataset.

by adding a params:pdb_paths and set the self.raw_dir to the root path(self.pdb_path) of pdb_paths list (the root path should be only one, pdb files should be under the same folder).

it allows loading pdb files from the self.pdb_path instead of loading from self.raw. If you wish to download from af2 or pdb, just set pdb_paths to None and it goes back to the former version.

CI

  • [Bugfix] - #208 explicitly installs jupyter_contrib_nbextensions in Docker.

1.5.1

Protein

  • [Feature] - #186 adds support for scaling node sizes in plots by a computed feature. Contribution by @cimranm
  • [Feature] - #189 adds support for parallelised download from the PDB.
  • [Feature] - #189 adds support for: van der waals interactions, vdw clashes, pi-stacking interactions, t_stacking interactions, backbone carbonyl-carbonyl interactions, salt bridges
  • [Feature] - #189 adds a residue_id column to PDB dfs to enable easier accounting in atom graphs.
  • [Feature] - #189 refactors torch geometric datasets to use parallelised download for faster dataset preparation.

Bugfixes

  • [Patch] - #187 updates sequence retrieval due to UniProt API changes.
  • [Patch] - #189 fixes bug where chains and PDB identifiers were not properly aligned in ml.ProteinGraphDataset.
  • [Patch] - #201 Adds missing MSE to graphein.protein.resi_atoms.RESI_NAMES, graphein.protein.resi_atoms.RESI_THREE_TO_1. #200
  • [Patch] - #201 Fixes bug where check for same-chain always evaluates as False. #199
  • [Patch] - #201 Fixes bug where deprotonation would only remove hydrogens based on atom_name rather than element_symbol. #198
  • [Patch] - #201 Fixes bug in ProteinGraphDataset input validation.

Breaking Changes

  • #189 refactors PDB download util. Now returns path to download file, does not accept a config object but instead receives the output directory path directly.

1.5.0

Protein

  • [Feature] - #165 adds support for direct AF2 graph construction.
  • [Feature] - #165 adds support for selecting model indices from PDB files.
  • [Feature] - #165 adds support for extracting interface subgraphs from complexes.
  • [Feature] - #165 adds support for computing the radius of gyration of a structure.
  • [Feature] - #165 adds support for adding distances to protein edges.
  • [Feature] - #165 adds support for fully connected edges in protein graphs.
  • [Feature] - #165 adds support for distance window-based edges for protein graphs.
  • [Feature] - #165 adds support for transformer-like positional encoding of protein sequences.
  • [Feature] - #165 adds support for plddt-like colouring of AF2 graphs
  • [Feature] - #165 adds support for plotting PyG Data object (e.g. for logging to WandB).
  • [Feature] - #170 Adds support for viewing edges in graphein.protein.visualisation.asteroid_plot. Contribution by @avivko.
  • [Patch] - #178 Fixes #171 and optimizes graphein.protein.features.nodes.dssp. Contribution by @avivko.
  • [Patch] - #174 prevents insertions always being removed. Resolves #173. Contribution by @OliverT1.
  • [Patch] - #165 Refactors HETATM selections.

Molecules

  • [Feature] - #165 adds additional graph-level molecule features.
  • [Feature] - #165 adds support for generating conformers (and 3D graphs) from SMILES inputs
  • [Feature] - #163 Adds support for molecule graph generation from an RDKit.Chem.Mol input.
  • [Feature] - #163 Adds support for multiprocess molecule graph construction.

RNA

  • [Feature] - #165 adds support for 3D RNA graph construction.
  • [Feature] - #165 adds support for generating RNA SS from sequence using the Nussinov Algorithm.

Changes

  • [Patch] - #163 uses tqdm.contrib.process_map insteap of multiprocessing.Pool.map to provide progress bars in multiprocessing.
  • [Fix] - #165 makes returned subgraphs editable objects rather than views
  • [Fix] - #165 fixes global logging set to "debug".
  • [Fix] - #165 uses rich progress for protein graph construction.
  • [Fix] - #165 sets saner default for node size in 3d plotly plots
  • [Dependency] - #165 Changes CLI to use rich-click instead of click for prettier formatting.
  • [Package] - #165 Adds support for logging with loguru and rich
  • [Package] - Pin BioPandas version to 0.4.1 to support additional parsing features.

Breaking Changes

  • #165 adds RNA SS edges into graphein.protein.edges.base_pairing
  • #163 changes separate filetype input paths to graphein.molecule.graphs.construct_graph. Interface is simplified to simply path="some/path.extension" instead of separate inputs like mol2_path=... and sdf_path=....

1.4.0 - UNRELEASED

  • [Patch] - #158 changes the eigenvector computation method from nx.eigenvector_centrality to nx.eigenvector_centrality_numpy.
  • [Feature] - #154 adds a way of checking that DSSP is executable before trying to use it. #154
  • [Feature] - #157 adds support for small molecule graphs using RDKit. Resolves #155.
  • [Feature] - #159 adds support for conversion to Jraph graphs for JAX users.

Breaking Changes

  • #157 refactors config matching operators from graphein.protein.config to graphein.utils.config
  • #157 refactors config parsing operators from graphein.utils.config to graphein.utils.config_parser

1.3.0 - 5/4/22

  • [Feature] - #141 adds edge construction based on sequence distance.
  • [Feature] - #143 adds equality and isomorphism testing functions between graphs, nodes and edges (#142)
  • [Feature] - #144 adds support for chain-level and secondary structure-level graphs with associated visualisation tools and tutorial. Resolves #128
  • [Feature] - #144 adds support for chord diagram visualisations.
  • [Feature] - #144 adds support for automagically downloading new PDB files for obsolete structures.
  • [Feature] - #150 adds support for hydrogen bond donor and acceptor counts node features. #145
  • [Misc] - #144 makes visualisation functions accessible in the graphein.protein namespace. #138
  • [Bugfix] - #147 fixes error in add_distance_threshold introduced in v1.2.1 that would prevent the edges being added to the graph. #146
  • [Bugfix] - #149 fixes a bug in add_beta_carbon_vector that would cause coordinates to be extracted for multiple positions if the residue has an altloc. Resolves #148

1.2.1 - 16/3/22

  • [Feature] - #124 adds support for vector features associated protein protein geometry. #120 #122
  • [Feature] - #124 adds visualisation of vector features in 3D graph plots.
  • [Feature] - #121 adds functions for saving graph data to PDB files.
  • [Bugfix] - #136 changes generator comprehension when updating coordinates in subgraphs to list comprehension to allow pickling
  • [Bugfix] - #136 fixes bug in edge construction functions using chain selections where nodes from unselected chains would be added to the graph.

Breaking Changes

  • #124 refactors graphein.protein.graphs.compute_rgroup_dataframe and moves it to graphein.protein.utils. All internal references have been moved accordingly.

1.2.0 - 4/3/2022

  • [Feature] - #104 adds support for asteroid plots and distance matrix visualisation.
  • [Feature] - #104 adds support for protein graph analytics (graphein.protein.analysis)
  • [Feature] - #110 adds support for secondary structure & surface-based subgraphs
  • [Feature] - #113 adds CLI support(!)
  • [Feature] - #116 adds support for onehot-encoded amino acid features as node attributes.
  • [Feature] - #119 Adds plotly-based visualisation for PPI Graphs
  • [Bugfix] - #110 fixes minor bug in asa where it would fail if added as a first/only dssp feature.
  • [Bugfix] - #110 Adds install for DSSP in Dockerfile
  • [Bugfix] - #110 Adds conda install & DSSP to tests
  • [Bugfix] - #119 Delaunay Triangulation computed over all atoms by default. Adds an option to restrict it to certain atom types.
  • [Bugfix] - #119 Minor fixes to stability of RNA Graph Plotting
  • [Bugfix] - #119 add tolerance parameter to add_atomic_edges
  • [Documentation] - #104 Adds notebooks for visualisation, RNA SS Graphs, protein graph analytics
  • [Documentation] - #119 Overhaul of docs & tutorial notebooks. Adds interactive plots to docs, improves docstrings, doc formatting, doc requirements.

Breaking Changes

  • #119 - Refactor RNA Graph constants from graphein.rna.graphs to graphein.rna.constants. Only problematic if constants were accessed directly. All internal references have been moved accordingly.

1.1.1 - 19/02/2022

  • [Bugfix] - #107 improves robustness of removing insertions and hetatms, resolves #98
  • [Packaging] - #108 fixes version mismatches in pytorch_geometric in docker install

1.1.0 - 19/02/2022

  • [Packaging] - #100 adds docker support.
  • [Feature] - #96 Adds support for extracting subgraphs
  • [Packaging] - #101 adds support for devcontainers for remote development.
  • [Bugfixes] - #95 adds improved robustness for edge construction functions in certain edge cases. Insertions in the PDB were occasionally not picked up due to a brittle implementations. Resolves #74 and #98

1.0.11 - 01/02/2022

  • [Improvement] - #79 Replaces Literal references with typing_extensions.Literal for Python 3.7 support.

1.0.10 - 23/12/2021

  • [Bug] Adds a fix for #74. Adding a disulfide bond to a protein with no disulphide bonds would fail. This was fixed by adding a check for the presence of a minimum of two CYS residues.