The k-mer features of a set of DNA/ Genomic sequences are of great significance on revealing the hidden patterns in that sequence population. Here, k is a integer constant and can range from 2 to several dozens, depending on the real application requirements. The k-mer features therefore are widely used in many applications of Bioinformatics, such as for building new prediction methods, etc.

This piece of Python code is for generating k-mer features for a list of DNA/ genomic sequences. Given a list of m DNA sequences, it returns a 2-d array with shape (m, 4^k) for the 1-hot representation of the kmer features. For a specific k, the total number of k-mer features is 4^k. For a DNA sequence, the value of each k-mer feature could be the number of occurrences of this k-mer, or its percentage of occurrences compared to all the other different k-mer features.

This code is carefully written to ensure efficiency. It runs sufficiently fast for a set of a very large number of sequences.

The code is written in Python 3.

How to use this code:

Example:

from kmer import kmer_featurization  # import the module kmer_featurization from the kmer.py file

seq_list = ['ATCGA', 'TCGAC']  # a list of DNA sequences

k = 6  # choose the value for k
obj = kmer_featurization(k)  # initialize a kmer_featurization object
kmer_features = obj.obtain_kmer_feature_for_a_list_of_sequences(seq_list, write_number_of_occurrences=False)
# If you would like the k-mer features to be the percentage of occurrences (ranging from 0 to 1) as stated above, then leave write_number_of_occurrences as False (the default). If you prefer the features to be the counts for each k-mer occurrence, then set it to True.

# If you just pass one sequence in string:
seq = 'ATCGAGC'
k = 6  
obj = kmer_featurization(k) 
kmer_feature = obj.obtain_kmer_feature_for_one_sequence(seq, write_number_of_occurrences=False)