Remove duplicated sequences from an alignment

The omit_duplicated app removes redundant sequences from a sequence collection (aligned or unaligned).

Let’s create sample data with duplicated sequences.

from cogent3 import make_aligned_seqs

data = {
    "a": "ACGT",
    "b": "ACG-",  # identical to 'a' except has a gap
    "c": "ACGG",  # duplicate
    "d": "ACGG",  # duplicate
    "e": "AGTC",  # unique
}
aln = make_aligned_seqs(data, moltype="dna")
aln

	0
a	ACGT
b	...-
c	...G
d	...G
e	.GTC

5 x 4 dna alignment

Creating the omit_duplicated app with the argument choose="longest" selects the duplicated sequence with the least number of gaps and ambiguous characters. In the above example, only one of c and d will be retained.

from cogent3 import get_app

omit_shorter_duplicate = get_app("omit_duplicated", moltype="dna", choose="longest")
omit_shorter_duplicate(aln)

	0
a	ACGT
b	...-
d	...G
e	.GTC

4 x 4 dna alignment

Creating the omit_duplicated app with the argument choose=None means only unique sequences are retained.

from cogent3 import get_app

omit_all_duplicates = get_app("omit_duplicated", moltype="dna", choose=None)
omit_all_duplicates(aln)

	0
a	ACGT
b	...-
e	.GTC

3 x 4 dna alignment

The mask_degen argument specifies how to treat matches between sequences with degenerate characters.

Let’s create sample data that has a DNA ambiguity code.

from cogent3 import make_aligned_seqs

aln = make_aligned_seqs(
    {
        "s1": "ATCG",
        "s2": "ATYG",  # matches s1 with ambiguity
        "s3": "GGTA",
    },
    moltype="dna",
)

Since “Y” represents pyrimidines where the site can be either “C” or “T”, s1 indeed matches s2 and one of them will be removed.

from cogent3 import get_app

app_dna = get_app("omit_duplicated", mask_degen=True, choose="longest", moltype="dna")
app_dna(aln)

	0
s1	ATCG
s3	GGTA

2 x 4 dna alignment