simScore {cosmo}R Documentation

Score motif detection simulation results

Description

This function computes sensitivity and specifity for the results returned by cosmo.

Usage

  simScore(truth, cosmoOut, minOverlap=0.25)

Arguments

truth align Alignment describing the true motif occurrences.
cosmoOut cosmo The results returned by cosmo().
minOverlap numeric A predicted motif must overlap at least this proportion of a known motif to be considered a hit.

Value

sens The proportion of true motif occurrences discovered (sensitivity).
spec The proportion of true motif occurrences among the discovered sites (specificity).
roc The area under the ROC curve.

Author(s)

Oliver Bembom, bembom@berkeley.edu

See Also

cosmo

Examples

## generate 20 sequences according to OOPS model
## with an expected 50
## motif
data(motifPWM)
data(transMats)
res <- rseq(20, 100, 1.0, motifPWM, transMats,"ZOOPS")
truth <- res$motifs
seqs <- res$seqs

res <- cosmo(seqs, constraints="None", minW=8, maxW=8)
simScore(truth, res)


[Package cosmo version 1.8.0 Index]