quad-table: Create and plot "quads" of samples

In almlab/texmexseq: Treatment Effect eXplorer for Microbial Ecology eXperiments (using Sequence Counts)

Description

Pull out a “quad” of samples from a larger data frame, then plot the pairs of samples in the quad against one another.

Usage

    quad.table(otu, control.before, control.after, treatment.before, treatment.after)
    quad.plot(quad)

Arguments

otu	data frame of a z- or F-transformed OTU table
control.before	name of the sample (in the OTU table) from the control unit before the treatment is applied to the treatment unit
control.after	sample from the control unit after treatment is applied
treatment.before	sample from the treatment unit before the treatment is applied
treatment.after	sample from the treatment unit after the treatment is applied
quad	a quad generated by quad.table

Details

texmexseq was designed to compare four samples at a time: two are “control” samples (before and after the treatment was applied to the “treatment” samples), the other two are the treatment samples.

quad.table will grab the columns with the four given names and make them into a new data frame (with OTU IDs kept as a separate column). This object can be plugged into quad.plot for viewing. quad.plot expects that the OTU table will be z- or F-transformed.

Value

quad.table	returns a data frame
quad.plot	returns a ggplot object

Author(s)

Scott Olesen [e-mail]

See Also

z.transform.table f.transform.table

Examples

# make up some data
sim.data <- function() rpoilog(1000, 1.0, 1.0, condS=TRUE)
otu <- data.frame(sample0=sim.data())
for (i in 1:10) otu[[paste('sample', i, sep='')]] <- sim.data()
otu.ids <- paste('otu', seq(1:1000), sep='')
rownames(otu) <- otu.ids
z.table <- z.transform.table(otu)

# pull out a quad, imagining that samples 1 and 2 were the control samples
# and 3 and 4 were the treatment
q <- quad.table(z.table, 'sample1', 'sample2', 'sample3', 'sample4')

# plot it
p <- quad.plot(q)
p

# ok, it's just a blob because we generated the data, but imagine we
# were particularly interested in OTUs that bloomed in the treatment
# but not in the control
interesting.otus <- filter(q, d.treatment > 2, d.control < 0)

# we can plot those in a different color
p + geom_point(data=interesting.otus, color='red')

# or see what their names are
head(arrange(interesting.otus, desc(d.treatment)))

almlab/texmexseq documentation built on May 10, 2019, 10:25 a.m.