How can we identify substrates that yield a positive reaction?

A Phenotype Microarray plate consists of 96 wells with up to 96 different substrates including positive and negative control. The physiological reactions towards these 96 substrates may be negative, weak, or positive.

This example demonstrates how to identify substrates that yield a positive reaction in the bacterial strain Escherichia coli DSM 18039.

Author: Johannes Sikorski

Load R packages and data

library(opm)
library(opmdata)
data(vaas_et_al)

The data set vaas_et_al contains data from four bacterial strains. For more details, see the respective publication.

Extract the plates data of strain Escherichia coli DSM 18039

DSM18039 <- vaas_et_al[c(Species = "Escherichia coli", Strain = "DSM18039",
  Experiment = "First replicate")]

DSM18039 consists of 10 plates:

dim(DSM18039)

## [1]  10 384  96

Discretise the maximum height (A) values into positive, weak, and

negative reactions

• The R package opm uses k-means partitioning to classify values into positive, weak, and negative reactions.
• Estimating weak reactions is optional at this step.
• There is also the concept of ambiguous reactions.
• There are other methods for discretisation.

DSM18039 <- do_disc(DSM18039)

Which wells contain substrates that yield positive reactions in all

plates?

wells(subset(DSM18039, positive = "all"))

##  [1] "A10" "A11" "A12" "B10" "B11" "B12" "C10" "C11" "D01" "D06" "D07"
## [12] "D09" "D10" "D11" "D12" "E02" "E03" "E05" "E09" "E10" "E11" "E12"
## [23] "F02" "F03" "F04" "F05" "F07" "F09" "F10" "F11" "F12" "G04" "G06"
## [34] "G07" "G08" "G11" "H03" "H07" "H08" "H10" "H11"

Textual listing of substrates that yield positive reactions

• Only substrates that yield positive reactions are listed. The output of weak or negative reactions is suppressed.

listing(subset(DSM18039, positive = "all"), ~ Strain)

## DSM18039
## --------
## Positive              Positive for positive control, pH 6, pH 5,
##                       1% NaCl, 4% NaCl, 8% NaCl, 1% sodium
##                       lactate, fusidic acid, D-sorbitol,
##                       D-glucose-6-phosphate,
##                       D-fructose-6-phosphate, D-serine #1,
##                       troleandomycin, rifamycin SV, minocycline,
##                       gly-pro, L-alanine, L-aspartic acid,
##                       L-serine, lincomycin, guanidine
##                       hydrochloride, niaproof, D-galacturonic
##                       acid, L-galactonic acid-gamma-lactone,
##                       D-gluconic acid, D-glucuronic acid, mucic
##                       acid, D-saccharic acid, vancomycin,
##                       tetrazolium violet, tetrazolium blue,
##                       L-lactic acid, alpha-keto-glutaric acid,
##                       D-malic acid, L-malic acid, lithium
##                       chloride, alpha-hydroxy-butyric acid,
##                       propionic acid, acetic acid, aztreonam and
##                       butyric acid.
## Negative              
## Ambiguous

Synopsis

• The R package opm contains functionality to discretise any of the four aggregated parameters either using k-means partitioning or other methods.
• The results can be retrieved in various forms, such as names of wells or names of substrates
• Additionally the results can be exported for phylogenetic analysis
• The results can be directly used for graphical display of either raw measurements or aggregated data, for example:

X-Y plot of substrates that yield positive reactions

The colour coding is according to the 10 different replicates.

xy_plot(subset(DSM18039, positive = "all"), include = list("Plate number"),
  neg.ctrl = 50, legend.fmt = list(space = "right"))