pandas-charm

pandas-charm is a small Python package for getting character matrices (alignments) into and out of pandas. Use this library to make pandas interoperable with BioPython and DendroPy.

Convert between the following objects:

• BioPython MultipleSeqAlignment <-> pandas DataFrame
• DendroPy CharacterMatrix <-> pandas DataFrame
• "Sequence dictionary" <-> pandas DataFrame

The code has been tested with Python 2.7, 3.5 and 3.6.

Source repository: https://github.com/jmenglund/pandas-charm

---

Table of contents

• Installation
• Running the tests
• Usage
  • DendroPy CharacterMatrix to pandas DataFrame
  • pandas DataFrame to Dendropy CharacterMatrix
  • BioPython MultipleSeqAlignment to pandas DataFrame
  • pandas DataFrame to BioPython MultipleSeqAlignment
  • "Sequence dictionary" to pandas DataFrame
  • pandas DataFrame to "sequence dictionary"
• The name
• License
• Citing
• Author

Installation

For most users, the easiest way is probably to install the latest version hosted on PyPI:

$ pip install pandas-charm

The project is hosted at https://github.com/jmenglund/pandas-charm and can also be installed using git:

$ git clone https://github.com/jmenglund/pandas-charm.git
$ cd pandas-charm
$ python setup.py install

You may consider installing pandas-charm and its required Python packages within a virtual environment in order to avoid cluttering your system's Python path. See for example the environment management system conda or the package virtualenv.

Running the tests

Testing is carried out with pytest:

$ pytest -v test_pandascharm.py

Test coverage can be calculated with Coverage.py using the following commands:

$ coverage run -m pytest
$ coverage report -m pandascharm.py

The code follow style conventions in PEP8, which can be checked with pycodestyle:

$ pycodestyle pandascharm.py test_pandascharm.py setup.py

Usage

The following examples show how to use pandas-charm. The examples are written with Python 3 code, but pandas-charm should work also with Python 2.7+. You need to install BioPython and/or DendroPy manually before you start:

$ pip install biopython
$ pip install dendropy

DendroPy CharacterMatrix to pandas DataFrame

>>> import pandas as pd
>>> import pandascharm as pc
>>> import dendropy
>>> dna_string = '3 5\nt1  TCCAA\nt2  TGCAA\nt3  TG-AA\n'
>>> print(dna_string)
3 5
t1  TCCAA
t2  TGCAA
t3  TG-AA

>>> matrix = dendropy.DnaCharacterMatrix.get(
...     data=dna_string, schema='phylip')
>>> df = pc.from_charmatrix(matrix)
>>> df
  t1 t2 t3
0  T  T  T
1  C  G  G
2  C  C  -
3  A  A  A
4  A  A  A

By default, characters are stored as rows and sequences as columns in the DataFrame. If you want rows to hold sequences, just transpose the matrix in pandas:

>>> df.transpose()
    0  1  2  3  4
t1  T  C  C  A  A
t2  T  G  C  A  A
t3  T  G  -  A  A

pandas DataFrame to Dendropy CharacterMatrix

>>> import pandas as pd
>>> import pandascharm as pc
>>> import dendropy
>>> df = pd.DataFrame({
...     't1': ['T', 'C', 'C', 'A', 'A'],
...     't2': ['T', 'G', 'C', 'A', 'A'],
...     't3': ['T', 'G', '-', 'A', 'A']})
>>> df
  t1 t2 t3
0  T  T  T
1  C  G  G
2  C  C  -
3  A  A  A
4  A  A  A

>>> matrix = pc.to_charmatrix(df, data_type='dna')
>>> print(matrix.as_string('phylip'))
3 5
t1  TCCAA
t2  TGCAA
t3  TG-AA

BioPython MultipleSeqAlignment to pandas DataFrame

>>> from io import StringIO
>>> import pandas as pd
>>> import pandascharm as pc
>>> from Bio import AlignIO
>>> dna_string = '3 5\nt1  TCCAA\nt2  TGCAA\nt3  TG-AA\n'
>>> f = StringIO(dna_string)  # make the string a file-like object
>>> alignment = AlignIO.read(f, 'phylip-relaxed')
>>> print(alignment)
SingleLetterAlphabet() alignment with 3 rows and 5 columns
TCCAA t1
TGCAA t2
TG-AA t3
>>> df = pc.from_bioalignment(alignment)
>>> df
  t1 t2 t3
0  T  T  T
1  C  G  G
2  C  C  -
3  A  A  A
4  A  A  A

pandas DataFrame to BioPython MultipleSeqAlignment

>>> import pandas as pd
>>> import pandascharm as pc
>>> import Bio
>>> df = pd.DataFrame({
...     't1': ['T', 'C', 'C', 'A', 'A'],
...     't2': ['T', 'G', 'C', 'A', 'A'],
...     't3': ['T', 'G', '-', 'A', 'A']})
>>> df
  t1 t2 t3
0  T  T  T
1  C  G  G
2  C  C  -
3  A  A  A
4  A  A  A

>>> alignment = pc.to_bioalignment(df, alphabet='generic_dna')
>>> print(alignment)
SingleLetterAlphabet() alignment with 3 rows and 5 columns
TCCAA t1
TGCAA t2
TG-AA t3

"Sequence dictionary" to pandas DataFrame

>>> import pandas as pd
>>> import pandascharm as pc
>>> d = {
...     't1': 'TCCAA',
...     't2': 'TGCAA',
...     't3': 'TG-AA'
... }
>>> df = pc.from_sequence_dict(d)
>>> df
  t1 t2 t3
0  T  T  T
1  C  G  G
2  C  C  -
3  A  A  A
4  A  A  A

pandas DataFrame to "sequence dictionary"

>>> import pandas as pd
>>> import pandascharm as pc
>>> df = pd.DataFrame({
...     't1': ['T', 'C', 'C', 'A', 'A'],
...     't2': ['T', 'G', 'C', 'A', 'A'],
...     't3': ['T', 'G', '-', 'A', 'A']})
>>> pc.to_sequence_dict(df)
{'t1': 'TCCAA', 't2': 'TGCAA', 't3': 'TG-AA'}

The name

pandas-charm got its name from the pandas library plus an acronym for CHARacter Matrix.

License

pandas-charm is distributed under the MIT license.

Citing

If you use results produced with this package in a scientific publication, please just mention the package name in the text and cite the Zenodo DOI of this project:

Choose your preferred citation style in the "Cite as" section on the Zenodo page.

Author

Markus Englund, orcid.org/0000-0003-1688-7112