dot-normForKmers: Adjust for k-mer composition (single iteration)

In fmicompbio/monaLisa: Binned Motif Enrichment Analysis and Visualization

Adjust for k-mer composition (single iteration)

Description

Adjust background sequence weights for differences in k-mer composition compared to the foreground sequences. This function implements a single iteration, and is called iteratively by .iterativeNormForKmers to get to the final set of adjusted weights, which will be the result of adjusting for GC and k-mer composition. The logic is based on Homer's normalizeSequenceIteration() function found in Motif2.cpp.

Usage

.normForKmers(
  kmerFreq,
  goodKmers,
  kmerRC,
  seqWgt,
  isForeground,
  minSeqWgt = 0.001,
  maxSeqWgt = 1000
)

Arguments

kmerFreq	a list with of matrices. The matrix at index i in the list contains the probability of k-mers of length i, for each k-mer (columns) and sequence (rows).
goodKmers	a list of numeric vectors; the element at index i contains the number of good (non-N-containing) k-mers of length i for each sequence.
kmerRC	a list of character vectors; the element at index i contains the reverse complement sequences of all k-mers of length i.
seqWgt	a numeric vector with starting sequence weights at the beginning of the iteration.
isForeground	logical vector of the same length as seqs. TRUE indicates that the sequence is from the foreground, FALSE that it is a background sequence.
minSeqWgt	Numeric scalar greater than zero giving the minimal weight of a sequence. The default value (0.001) is based on Homer (HOMER_MINIMUM_SEQ_WEIGHT constant in Motif2.h).
maxSeqWgt	Numeric scalar greater than zero giving the maximal weight of a sequence. The default value (1000) is based on HOMER (1 / HOMER_MINIMUM_SEQ_WEIGHT constant in Motif2.h).

Value

a named list with elements seqWgt (updated weights) and err (error measuring difference of foreground and weighted background sequence compositions).

fmicompbio/monaLisa documentation built on July 10, 2024, 8:44 a.m.