findSeedMatches: Predicting and characterizing miRNA binding sites

View source: R/scanning.R

findSeedMatchesR Documentation

Predicting and characterizing miRNA binding sites

Description

'findSeedMatches' takes a set of sequences and a set of miRNAs (given either as target seeds, mature miRNA sequences, or a KdModelList).

Usage

findSeedMatches(
  seqs,
  seeds,
  shadow = 0L,
  onlyCanonical = FALSE,
  maxLogKd = c(-1, -1.5),
  keepMatchSeq = FALSE,
  minDist = 7L,
  p3.extra = FALSE,
  p3.params = list(maxMirLoop = 7L, maxTargetLoop = 9L, maxLoopDiff = 4L, mismatch =
    TRUE, GUwob = TRUE),
  agg.params = .defaultAggParams(),
  ret = c("GRanges", "data.frame", "aggregated"),
  BP = NULL,
  verbose = NULL,
  n_seeds = NULL,
  useTmpFiles = FALSE,
  keepTmpFiles = FALSE
)

Arguments

seqs

A character vector or 'DNAStringSet' of DNA sequences in which to look.

seeds

A character vector of 7-nt seeds to look for. If RNA, will be reversed and complemented before matching. If DNA, they are assumed to be the target sequence to look for. Alternatively, a list of objects of class 'KdModel' or an object of class 'KdModelList' can be given.

shadow

Integer giving the shadow, i.e. the number of nucleotides hidden at the beginning of the sequence (default 0).

onlyCanonical

Logical; whether to restrict the search only to canonical binding sites.

maxLogKd

Maximum log_kd value to keep. This has a major impact on the number of sites returned, and hence on the memory requirements. Set to Inf to disable (_not_ recommended when running large scans!).

keepMatchSeq

Logical; whether to keep the sequence (including flanking dinucleotides) for each seed match (default FALSE).

minDist

Integer specifying the minimum distance between matches of the same miRNA (default 7). Closer matches will be reduced to the highest-affinity. To disable the removal of overlapping features, use 'minDist=-Inf'.

p3.extra

Logical; whether to keep extra information about 3' alignment. Disable (default) this when running large scans, otherwise you might hit your system's memory limits.

p3.params

Named list of parameters for 3' alignment with slots 'maxMirLoop' (integer, default = 7), 'maxTargetLoop' (integer, default = 9), 'maxLoopDiff' (integer, default = 4), 'mismatch' (logical, default = TRUE) and 'GUwob' (logical, default = TRUE).

agg.params

A named list with slots 'a', 'b', 'c', 'p3', 'coef_utr', 'coef_orf' and 'keepSiteInfo' indicating the parameters for the aggregation. Ignored if 'ret!="aggregated"'. For further details see documentation of 'aggregateMatches'.

ret

The type of data to return, either "GRanges" (default), "data.frame", or "aggregated" (aggregates affinities/sites for each seed-transcript pair).

BP

Pass 'BiocParallel::MulticoreParam(ncores, progressbar=TRUE)' to enable multithreading.

verbose

Logical; whether to print additional progress messages (default on if not multithreading)

n_seeds

Integer; the number of seeds that are processed in parallel to avoid memory issues.

useTmpFiles

Logical; whether to write results for single miRNAs in temporary files (ignored when scanning for a single seed). Alternatively, 'useTmpFiles' can be a character vector of length 1 indicating the path to the directory in which to write temporary files.

keepTmpFiles

Logical; whether to keep the temporary files at the end of the process; ignored if 'useTmpFiles=FALSE'. Temporary files are removed only upon successful completion of the function, meaning that they will not be deleted in case of errors.

Value

A GRanges of all matches. If 'seeds' is a 'KdModel' or 'KdModelList', the 'log_kd' column will report the ln(Kd) multiplied by 1000, rounded and saved as an integer. If 'ret!="GRanges', returns a data.frame.

Examples

# we create mock RNA sequences and seeds:
seqs <- getRandomSeq(n=10)
seeds <- c("AAACCAC", "AAACCUU")
findSeedMatches(seqs, seeds)

ETHZ-INS/scanMiR documentation built on July 17, 2024, 6:18 a.m.