This constraint is open source. Any third-party models, product names, or trademarks referenced are the property of their respective owners, and Proto is not affiliated with them.
evo-design/proto-language/proto_language/constraint/sequence_composition/dinucleotide_composition_constraint.pyConstraint contributors
adititm
Score the distance between observed and reference dinucleotide composition.
The observed overlapping-dinucleotide frequency distribution of each sequence
is compared to adititmconfig.reference_frequencies using the configured distance
metric. Lower scores mean the sequence’s nucleotide-pair usage is closer to the
reference (natural) composition.
API Reference
Configuration for the dinucleotide-composition match constraint.Scores the distance between a sequence’s observed dinucleotide-frequency
distribution and a reference distribution. The reference is supplied as a
mapping of dinucleotide -> frequency (e.g. derived from natural sequences);
RNA (
U) keys are folded onto the DNA alphabet (T) and the reference is
renormalized to sum to 1.object
required
Target dinucleotide -> frequency mapping (natural composition profile).
enum
default:"total_variation"
Distribution distance: ‘total_variation’ (0.5*sum|p-q|) or normalized ‘l2’.Options:
total_variation, l2number
default:"1.0"
Multiplier applied to the raw distance before clamping to [0, 1].
ReturnsConstraintOutput
One result per sequence. score is in [0, 1]
(0.0 = observed composition matches the reference). metadata carries
dinucleotide_distance (the raw, pre-scale distance) and
dinucleotide_frequencies (the observed distribution, or None when
the sequence is too short).Usage
python