get_wis: WIS-algorithm for a set of intervals

In robbueck/wisard: Perform wighted interval scheduling on a Genomic Ranges object

Description

Performs the WIS algorithm on a set of weighted intervals given as a GRangesList object.

Usage

get_wis(
  granges_list,
  overlap = 0,
  max_score = "score",
  score_funct = function(x, y, n) x + y,
  max_funct = function(x, y, n) x + y
)

Arguments

granges_list	A set of intervals as a GRangesList object. Requires an interval-score to maximize over as a column specified in raw_score.
overlap	A non-negative integer. The allowed overlap between two intervals.
max_score	Name of a column in granges_list containing integers or floats to be used as an interval-score to maximize over.
score_funct	Function to update the score of a set each time an element is added. Can take up to three arguments: x (set-score of set so far) y (interval-score of added item) and n (number of items in set including the one added). Returnes set-score, default: x + y.
max_funct	Function to calculate a set-score to maximize over. This set-score is not used to calculate itself, but is calculated using the set-score from score_funct. Takes up to three arguments: x (set-score of set so far), y (interval-score of added item), n (number of items in set including the one added). Returns set-score, default: x + y. For initial calculation y = 0 and n = 2. If in doubt about its use, set equal to score_funct.

Details

This function calculates a set of non-overlapping intervals while maximizing over a given set-score. The behaviour of this set-score when an element is added to a list is specified by the two input parameters score_funct and max_funct. For a simple sum or product set-score, set both functions to x + y or x * y.

The max_score option can be used, if a set-score contains non linear elements, such as log(n), e.g. in the BLAST scum-score:

score_funct = function (x, y, n) (x + y)

max_funct = function (x, y, n) (x + y) + log(n!)

If n is included in one of the formulars in a non-linear way, it is not guaranteed anymore that the final set is the best possible set.

Strand information is ignored here, has to be taken into account before calling this fucntion.

Value

A GRangesList object with intervals compatible according to the specified options.

robbueck/wisard documentation built on Jan. 25, 2022, 12:35 a.m.