Installation

With npm:

npm install @rnacanvas/position-pairs

Usage

All exports of this package can be accessed as named imports.

// some example imports
import type { PositionPair } from '@rnacanvas/position-pairs';
import { parseDotBracket } from '@rnacanvas/position-pairs';

PositionPair

The PositionPair type represents a pairing between two positions in a sequence.

Positions in a sequence use one-based indexing (as opposed to zero-based indexing).

Thus, the first item in a sequence will have a position of 1, and so on.

// the two numbers should be positive integers
type PositionPair = [number, number];

Code in this package will also generally handle the case of a position paired with itself (i.e., a position pair composed of the same position twice), though such a pairing would not have any biological underpinning.

deepCopy()

Creates and returns a deep copy of an individual position pair or an array of position pairs.

deepCopy([82, 108]); // [82, 108]

deepCopy([[19, 2], [10, 30], [11, 29]]); // [[19, 2], [10, 30], [11, 29]]

areEqual()

Returns true if and only if two pairs pair the same two positions together.

areEqual([11, 22], [11, 22]); // true
areEqual([11, 22], [22, 11]); // true

areEqual([11, 22], [12, 21]); // false
areEqual([11, 22], [11, 25]); // false

This function also handles the case of a position paired with itself.

areEqual([11, 11], [11, 11]); // true

areEqual([11, 11], [12, 12]); // false
areEqual([11, 11], [11, 22]); // false

areConflicting()

Returns true if and only if two pairs conflict with one another (i.e., they both share a position but they pair it with two different positions).

areConflicting([10, 20], [10, 21]); // true
areConflicting([10, 20], [9, 20]); // true

areConflicting([10, 20], [11, 21]); // false

// the two pairs are equal
areConflicting([10, 20], [10, 20]); // false

areCrossing()

Returns true if and only if two pairs form a pseudoknot together.

This would correspond to the lines of the two pairs crossing each other in the circle diagram of a nucleic acid structure.

areCrossing([10, 30], [20, 40]); // true
areCrossing([40, 20], [30, 10]); // true

areCrossing([10, 39], [11, 38]); // false

// are equal
areCrossing([22, 64], [22, 64]); // false

areKnotless()

Returns true if and only if the given pairs do not form any pseudoknots.

// a simple stem
var pairs = [[3, 24], [4, 23], [5, 22], [6, 21]];

areKnotless(pairs); // true

// add an H-type pseudoknot
pairs.push([9, 40], [10, 39], [11, 38], [12, 37]);

areKnotless(pairs); // false

knotless()

Creates and returns an entirely new array of pairs, with pairs creating pseudoknots having been omitted.

Pairs are omitted/retained according to the incremental range heuristic reported by Smit et al., 2008.

// form an H-type pseudoknot
var pairs = [
  [7, 18], [8, 17], [9, 16],
  [11, 25], [12, 24], [13, 23],
];

knotless(pairs); // [[7, 18], [8, 17], [9, 16]]

Structure

The Structure type is a way to express a nucleic acid structure using position pairs.

// a sequence of nucleobases, for instance
type Sequence<T> = T[];

// a sequence and pairs among the items in the sequence
type Structure<T> = [Sequence<T>, PositionPair[]];

mountainPlotTraversal()

The mountainPlotTraversal() function returns an array of numbers that is the mountain plot traversal of a given structure (i.e., each value in the returned array indicates the height of the mountain plot at that position in the structure).

Mountain plot heights are calculated using the ViennaRNA definition.

This function will throw if the passed in structure contains pseudoknots.

var seq = [...'1234567890ab'];
var pairs = [[2, 11], [3, 10], [4, 9]];

var mpt = mountainPlotTraversal(seq, pairs);

mpt; // [0, 0, 1, 2, 3, 3, 3, 3, 2, 1, 0, 0]