Logomaker

Common errors

• AttributeError: module 'logomaker' has no attribute 'Logo'

  cause You are trying to access the `Logo` class using `logomaker.Logo` after importing the module as `import logomaker`, but the `Logo` class is intended for direct import.
  fix Change your import statement from `import logomaker` to `from logomaker import Logo`, and then use `Logo(...)` directly.

• ValueError: Dataframe must have an index of positive integers.

  cause The pandas DataFrame provided to the `logomaker.Logo` constructor (or related functions) does not have a numeric index, or the index contains non-positive integers. Logomaker expects the index to represent sequence positions.
  fix Ensure your input DataFrame has a numeric index (e.g., `0, 1, 2...` or `1, 2, 3...`). You might need to use `df.reset_index(drop=True)` or `df.set_index(df.index + 1)` if your index is not suitable.

• KeyError: 'A'

  cause Your input DataFrame (e.g., a position-weight matrix) is missing columns for expected character labels (e.g., 'A', 'C', 'G', 'T' for DNA/RNA). Logomaker expects columns corresponding to the characters you want to represent.
  fix Ensure your input pandas DataFrame has columns for all the characters you intend to display in the logo. For DNA/RNA, these are typically 'A', 'C', 'G', 'T' (or 'U'). For proteins, this would be the 20 standard amino acid codes.

Warnings

• gotcha Older versions (pre-0.8.2) frequently rebuilt the Matplotlib font cache when loading, leading to slow startup times and potential resource issues, especially in scripts or automated environments.
  Fix: Upgrade to logomaker version 0.8.2 or newer to benefit from performance improvements related to Matplotlib caching.
• gotcha The `stack_order='fixed'` option for `Logo` objects was non-functional in versions prior to 0.8.2. This could lead to incorrect rendering when trying to maintain a specific, fixed order of characters in the logo stacks.
  Fix: Upgrade to logomaker version 0.8.2 or newer to ensure the `stack_order='fixed'` option works as intended.
• gotcha Versions prior to 0.8.3 had overly stringent numerical type checking for input matrices. This could cause valid input DataFrames to be rejected if their numerical types (e.g., `int64` vs. `float64`) did not precisely match internal expectations.
  Fix: Upgrade to logomaker version 0.8.3 or newer. Alternatively, explicitly cast your input DataFrame columns to `float64` or `int64` as appropriate before passing them to Logomaker functions.

Install

• pip install logomaker Install Logomaker

Imports

• Logo
  wrong:

  import logomaker; logomaker.Logo

  correct:

  from logomaker import Logo

  The Logo class is typically imported directly for clarity and common usage patterns.
• get_example_matrix
  wrong:

  import logomaker; logomaker.get_example_matrix

  correct:

  from logomaker import get_example_matrix

  Utility function for loading example data, usually imported directly.

Quickstart

This quickstart demonstrates how to load an example sequence matrix, create a `Logo` object, and apply basic styling before displaying the logo using Matplotlib.

import logomaker
import matplotlib.pyplot as plt
import pandas as pd

# Load an example matrix (e.g., position-weight matrix)
# This returns a pandas DataFrame
example_df = logomaker.get_example_matrix('ww_matrix')

# Create a Logo object from the DataFrame
logo = logomaker.Logo(example_df)

# Customize the logo (optional)
logo.style_spines(visible=False)
logo.style_xticks(rotation=90, fmt='%d', anchor=0)
logo.highlight_position(p=6, color='gold')
logo.ax.set_ylabel('Information (bits)')
logo.ax.set_xlabel('Position')
logo.ax.set_title('RNA Binding Protein WW Domain Logo')

plt.tight_layout()
plt.show()