cellOrigins (version 0.1.3)

discovery_probability: In situ discovery probability as a function of FPKM

Description

Groups transcripts by expression strength and calculates for each such group the percentage of genes that gave a positive staining signal in the in situ hybridisation.

If the sequenced material matches the in situ hybridisation tissue, then weakly expressed genes in the sequenced material should be rearely in the in situ staining set of genes. Strongly expressed genes should correspondingly often also stain during hybridisation. Overall, if the match is not spurious, there should be a logarithmic dose-response relationship between sequencing read coverage and staining probability. In a plot of discovery probability against log(coverage) this shows as an approximately straight line (see example).

Usage

discovery_probability(seq_signature, terms, cut.points,
    insitu=cellOrigins::BDGP_insitu_dmel_embryo)

Arguments

seq_signature

A named vector containing FPKM RNAseq data. Each element name must correspond to the names used in the insitu argument. NAs are permitted.

terms

A vector of anatomical terms which together are assumed to be the origin of the RNAseq data.

cut.points

A vector of cut points for grouping of values. E.g. 0:3 denotes the bins 0<=x<1, 1<=x<2, 2<=x<3.

insitu

Matrix with in situ hybridisation data. Rows are transcript names (same names as used for seq_signature) and coloumns are anatomical terms (possibly combined with developmental stages). 1 denotes staining of a particular transcript in a particular tissue, 0 denotes no staining. Defaults to BDGP_insitu_dmel_embryo, a staining dataset for Drosophila melanogaster embryos.

Value

A matrix with a row for each bin and three coloumns. The first coloumn is the probability of discovery, the second the number of transcripts in the expression bin that were discovered by in situ hybridisation. The third coloumn is the total number of transcripts in the bin.

See Also

iterating_seqVsInsitu, BDGP_insitu_dmel_embryo, discovery.log, discovery.linear, discovery.identic, prior.temporal_proximity_is_good, prior.all_equal, diagnosticPlots.

Examples

# NOT RUN {
fpath <- system.file("extdata", "vncMedianCoverage.tsv", package="cellOrigins")
vncExpression <- read.delim(file = fpath, header=FALSE, as.is=TRUE)

expression <- vncExpression$V2
names(expression) <- vncExpression$V1

p <- discovery_probability(expression,
  "6|ventral nerve cord", c(0, 2^(0:10)))

plot(x=-1:9, y=p[,1], type="l",
  xlab="log2(FPKM)", ylab="p(discovery in situ)")

# }