Remove problem sequences from an alignment

Using omit_bad_seqs we can eliminate sequences from an Alignment based on their gap fraction and/or the number of gaps they uniquely introduce.

Let’s create a sample alignment with some gaps.

from cogent3 import make_aligned_seqs

aln = make_aligned_seqs(
    {
        "s1": "---ACC---TT-",
        "s2": "---ACC---TT-",
        "s3": "---ACC---TT-",
        "s4": "--AACCG-GTT-",
        "s5": "--AACCGGGTTT",
        "s6": "AGAACCGGGTT-",
        "s7": "------------",
    },
    moltype="dna",
)

Removing sequences with more than X% gaps

Creating the omit_bad_seqs app with the argument gap_fraction=0.5 will omit sequences that contain 50% or more gaps.

from cogent3 import get_app

omit_frac_05 = get_app("omit_bad_seqs", gap_fraction=0.5)
omit_frac_05(aln)

	0
s6	AGAACCGGGTT-
s4	--.....-....
s5	--.........T

3 x 12 dna alignment

Removing sequences that contribute many gaps

The quantile=0.8 argument omits sequences that are ranked above the specified quantile with respect to the number of gaps uniquely introduced into the alignment. In the following example, sequence s6 is omitted, as it uniquely introduces gaps in the first two positions of the alignment.

from cogent3 import get_app

omit_quant_08 = get_app("omit_bad_seqs", quantile=0.8)
omit_quant_08(aln)

	0
s5	--AACCGGGTTT
s1	..-...---..-
s2	..-...---..-
s3	..-...---..-
s4	.......-...-

5 x 12 dna alignment