Alphabets

`CharAlphabet` and `MolType`

MolType instances have an CharAlphabet.

from cogent3 import get_moltype

dna = get_moltype("dna")
print(dna.alphabet)
protein = get_moltype("protein")
print(protein.alphabet)

('T', 'C', 'A', 'G')
('A', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'K', 'L', 'M', 'N', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'Y')

CharAlphabet instances reference the MolType that created them.

dna.alphabet.moltype is dna

True

Creating tuple alphabets

You can create a tuple alphabet of, for example, dinucleotides or trinucleotides.

dinuc_alphabet = dna.alphabet.get_kmer_alphabet(2)
trinuc_alphabet = dna.alphabet.get_kmer_alphabet(3)
print(dinuc_alphabet, trinuc_alphabet, sep="\n")

('TT', 'TC', 'TA', 'TG', 'CT', 'CC', 'CA', 'CG', 'AT', 'AC', 'AA', 'AG', 'GT', 'GC', 'GA', 'GG')
('TTT', 'TTC', 'TTA', 'TTG', 'TCT', 'TCC', 'TCA', 'TCG', 'TAT', 'TAC', 'TAA', 'TAG', 'TGT', 'TGC', 'TGA', 'TGG', 'CTT', 'CTC', 'CTA', 'CTG', 'CCT', 'CCC', 'CCA', 'CCG', 'CAT', 'CAC', 'CAA', 'CAG', 'CGT', 'CGC', 'CGA', 'CGG', 'ATT', 'ATC', 'ATA', 'ATG', 'ACT', 'ACC', 'ACA', 'ACG', 'AAT', 'AAC', 'AAA', 'AAG', 'AGT', 'AGC', 'AGA', 'AGG', 'GTT', 'GTC', 'GTA', 'GTG', 'GCT', 'GCC', 'GCA', 'GCG', 'GAT', 'GAC', 'GAA', 'GAG', 'GGT', 'GGC', 'GGA', 'GGG')

Convert a sequence into integers

seq = "TAGT"
indices = dna.alphabet.to_indices(seq)
indices

array([0, 2, 3, 0], dtype=uint8)

Convert integers to a sequence

import numpy

data = numpy.array([0, 2, 3, 0], dtype=numpy.uint8)
seq = dna.alphabet.from_indices(data)
seq

'TAGT'

Converting a sequence into k-mer indices

You can use a KmerAlphabet to convert a standard sequence into a numpy array of integers. In this case, each integer is the encoding of the dinucleotide string into the index of that dinucleotide. Because the CharAlphabet and KmerAlphabet both inherit from tuple, they have the built-in .index() method.

import numpy

bases = dna.alphabet
dinucs = bases.get_kmer_alphabet(2)
dinucs.index("TG")

The to_indices() method is faster and provides more flexibility. We use that on the single dinucleotide

dinucs.to_indices("TG")

array([3], dtype=uint8)

and on a longer sequence where we want the independent k-mers.

seq = "TGTGGCACAAATACTCATGCCAGCTCATTA"
dinuc_indices = dinucs.to_indices(seq, independent_kmer=True)
dinuc_indices

array([ 3,  3, 13,  9, 10,  8,  9,  1,  8, 13,  6, 13,  1,  8,  2],
      dtype=uint8)

We can also convert the sequence into all possible k-mers.

dinucs.to_indices(seq, independent_kmer=False)

array([ 3, 12,  3, 15, 13,  6,  9,  6, 10, 10,  8,  2,  9,  4,  1,  6,  8,
        3, 13,  5,  6, 11, 13,  4,  1,  6,  8,  0,  2], dtype=uint8)

Quality score converters

make_qual_converter builds a callable that maps a fastq quality string (as bytes) into a numpy.uint8 array of Phred scores. Both Phred+33 and Phred+64 encodings are supported.

from cogent3.core.alphabet import make_qual_converter

conv = make_qual_converter("phred+33")
conv(b"!I~")

array([ 0, 40, 93], dtype=uint8)

See Directly use the genbank format parser to load a sequence and annotations for an example combining this with iter_fastq_records.

CharAlphabet and MolType