discovery.log {cellOrigins}

Calculates discovery probability by RNA in situ hybridisation given a sequencing signal

Description

A set of functions with different assumptions on the probability of RNA in situ staining, given a sequencing coverage.

Usage

discovery.log(seq, saturate = 60, bias = 0.01)
discovery.linear(seq, saturate = 60, bias = 0.01)
discovery.identic(seq, saturate=Inf, bias=0)

Arguments

seq

A vector of sequencing FPKMs.

saturate

FPKM value from which on maximum discovery probability (=0.99) is assumed (i.e. almost certain true positives). Value of 60 is default, may need adjustment to sequencing coverage.

bias

Positive staining probability of 0 FPKM transcripts (i.e. false positives). Must be >0. Default is 0.01, an empirically determined value.

Details

1. discovery.log Uses a logarithmic saturation function for discovery probabilities. This relationship was empirically determined from sequencing and hybridisation data.

2. discovery.linear Linear saturation function for discovery probabilities.

3. discovery.identic Passes input through. Useful for comparing RNASeq Vs. RNASeq data. Also for cases when the discovery probability for each transcript has been already determined in some other way.

Value

A vector of probabilities. Element names are preserved.

See Also

seqVsInsitu

Examples

plot(0:80, discovery.log(0:80),
  ylim=c(0,1.1), type="l",
  xlab="FPKM", ylab="p(discovery insitu hybridization)")

plot(0:80, discovery.linear(0:80),
  ylim=c(0,1.1), type="l",
  xlab="FPKM", ylab="p(discovery insitu hybridization)")