Position Weight Matrix
-------------------------

This tutorial shows how to use rollable functions. Reading a motif-pwm from file, a PositionWeightMatrix function is created using the appropriate alphabet and counts. Since `PositionWeightMatrix` is a `RollableFunction` subclass it has a rolling_window method that applies the pwm to all valid windows in the sequence set.

.. code-block:: python

    import bionumpy as bnp
    import matplotlib.pyplot as plt
    from bionumpy.jaspar import read_jaspar_matrix
    from bionumpy.position_weight_matrix import PositionWeightMatrix, pwm_from_counts
    from bionumpy.encodings.alphabet_encoding import get_alphabet_array_class
    
    # Read the alphabet and counts from jaspar file
    alphabet, matrix = read_jaspar_matrix("example_data/MA0080.1.jaspar")
    
    # Convert counts to position weight matrix
    pwm = pwm_from_counts(matrix)
    
    # Make an array-class for the alphabet
    arrayclass = get_alphabet_array_class(alphabet)
    
    # Get the motif score function
    motif_score = PositionWeightMatrix(pwm, arrayclass)
    
    #Get reads
    entries = bnp.open("example_data/big.fq.gz").read()
    
    # Calculate the motif score for each valid window
    scores = motif_score.rolling_window(entries.sequence)
    
    # Get a histogram of the max-score for each read
    plt.hist(scores.max(axis=-1)); plt.show()