Biotite

Changelog

Additions

- Added MMTF writing functionality
- Added `set_array()` method for `structure.io.mmtf.MMTFFile` allows for array encoding
- `write()` method for `structure.io.mmtf.MMTFFile` is now implemented
- `structure.io.mmtf.set_structure()` is used for writing an
`structure.AtomArray` or `structure.AtomArrayStack` instance
to the fields of an `structure.io.mmtf.MMTFFile` instance
- Added `structure.get_residue_starts()`
- Returns the indices of residue starts in an atom array (stack)
- Internally used by all other residue level functions, simplifying the code base
- Added `structure.BondList`
- Stores indices and bond types of bonded atoms in a corresponding atom array (stack)
- Can be associated to an atom array (stack) using the `bonds` attribute
- Indexing behaves in accordance to the atom array (stack)
- C-accelerated operations
- Can be loaded from and written to an MMTF file
- Added convenience function `structure.io.save_structure()`
- Added visualization capabilities based on `matplotlib` figures
- Using `Visualizer` as base class
- `sequence.graphics.FeatureMap` for visualization of sequence annotations
- `sequence.graphics.SequenceLogo` for creation of sequence logos
- `sequence.graphics.AlignmentVisualizer` for displaying alignments
- `sequence.graphics.AlignmentSimilarityVisualizer` colors symbols
based on similarities in alignment column
- `sequence.graphics.AlignmentSymbolVisualizer` colors symbols
based on a color scheme
- Color schemes that can be loaded from JSON files
- Biotite standard colors via `colors` list
- Added support of NCBI Entrez search service
- `database.entrez.Query` instances that can be combined using boolean operators
- `database.entrez.Query` objects a re submitted via `database.entrez.search()`
- Added `sequence.io.genbank.MultiFile` for Genbank or Genpept files with multiple contents
- Added `get_order()` method to `application.MSAApp`
- Added PEP 8 compliant `__author__` attribute for each module
- Added example gallery to documentation

Changes

- Higher performance of `structure.superimpose()`
- Renamed `structure.io.get_structure from()` to `structure.io.load_structure()`
- `sequence.Alphabet` objects are hashable

Fixes
- Superimposition was not working for `structure.AtomArrayStack` instances
with more models than atoms
- Input value checks were missing in `structure.AdjacencyMap` constructor
- `box_size` parameter in `structure.AdjacencyMap` constructor
was restricted to integer values
- alphabet checks in `application.MSAApp` constructor were missing
- `\\` was required at the end of Genbank files
- Index arrays did not work as second index for `sequence.align.Alignment` objects
- `Enum` types are now properly documented

Notes
Currently no wheels are provided for Python 3.5 on Windows due to issues with `openssl` in the install process.

Changelog

- Slightly higher performance of `sasa()`
- Fixed issue with `sasa()` in combination with Windows x64 version

Changelog

- *matplotlib* is now an optional requirement
- Added `mass_of_element()`, `atom_masses()`, `mass_center()` and `gyration_radius()` to `structure` subpackage
- Increased performance of `AdjacencyMap` and `sasa()`

Note
Currently the build fails for the Windows x64 wheels, due to an error in the unit tests for the new SASA implementation. This will (hopefully) be fixed in the next version.

Changelog

- Added ``get_residue_count()`` in ``structure`` subpackage
- Added ``get_symbol_frequency()`` method to ``Sequence``
- Added ``terminal_penalty`` parameter to ``align_optimal()``
- If set to false, terminal gaps are not penalized
- Fixed parser for trajectory files

Changelog

- Removed pure Python implementation
- Added wheels for Windows
- Added support for sequence features
- Added GenBank and GenPept parser
- A ``Feature`` describes a functional part of a sequence
- An ``Annotation`` is a collection of ``Feature`` objects corresponding to one sequence
- An ``AnnotatedSequence`` combines an ``Annotation``with a ``Sequence``
- More indexing options for ``Sequence`` objects
- Moved static method ``Sequence.encode()`` and ``Sequence.decode()``
to alphabet member functions ``encode_multiple()`` and ``decode_multiple()``
- Added ``Alphabet`` subclass ``LetterAlphabet``
- ``LetterAlphabet`` has massively improved performance for encoding and decoding
- Removed ``is_letter_alphabet()`` method from ``Alphabet``
- Improved performance of ``sequence.align.simple_score()``
- ``SubstitutionMatrix.score_matrix()`` does not copy the internal ``ndarray``
- Added 'MSE' and 'SEC' as 3-letter amino acids in ``ProteinSequence``
- ``ProteinSequence`` accepts lower case strings in constructor
- Improved PDB file support
- Ability to choose insertion codes and alternative locations
- Added model parameter
- Added MMTF file parser
- ``MMTFFile`` instances efficiently decode MMTF file contents
- ``biotite.structure.io.mmtf.get_structure()`` constructs ``AtomArray`` and ``AtomArrayStack``
instances from ``MMTFFile``
- ``biotite.structure.io.get_model_from()`` accepts MMTF files
- Added ``DSSPApp`` to interface DSSP for secondary structure annotation
- Added ``MafftApp`` and ``ClustalOmegaApp`` for multiple sequence alignments

Changelog

- Fixed bug in coordinate superimposition that occured when the rotation matrix contained a reflection