In PieGes/synerdrug: Analysis of drug combinations

require(knitr)
opts_chunk$set(echo = TRUE, fig.width = 5, fig.height = 4, prompt = TRUE, warning = FALSE)

require(synerdrug)
data(exDrugSyn)

exDrugSyn
drugNames(exDrugSyn)
doses(exDrugSyn)

Visualise data

The data can be visualized by using plot on the DrugSyn object. Several visualization types are available: as heatmaps, as surface or parallel lines.

plot(exDrugSyn)

The data can be viewed as a surface by setting the type argument of plot to "surface". Arguments from persp can be passed to plot to adjust the appearance (e.g. azimuth and co-latitude).

plot(exDrugSyn, type = "surface")
plot(exDrugSyn, type = "surface", theta = -45, phi = 45)

plot(exDrugSyn, type = "parallel", ref = "drugA")
plot(exDrugSyn, type = "parallel", ref = "drugB")

Individual responses

The single agent responses are fitted by a Hill function (also known as log-logistic model or sigmoid curve) with 4 parameters:

$$E=E_0 + \frac{E_{max}-E_0}{1 + (\frac{EC_{50}}{d})^n}$$

where:

• $E$ is the observed effect
• $d$ is the dose
• $E_0$ is the minimum effect
• $E_{max}$ is the maximum effect
• $EC_{50}$ is the effective dose 50
• $n$ is the sigmoidicity parameter.

plot(exDrugSyn, type = "ind", drug = "drugA")
plot(exDrugSyn, type = "ind", drug = "drugB")

The parameters of the hill functions can be accessed with respInd.

respInd(exDrugSyn)

Synergy estimates

Several synergy estimate idexes are computed by synerdrug.

Effect-based synergy evaluation

$HSA_{ab}=max(E_{a0}, E_{0b})/E_{ab}$;

$Bliss_{ab}=(E_{a0}+E_{0b}-E_{a0}E_{0b})/E_{ab}$

plot(exDrugSyn, type = "heatmap", what = "Bliss")
plot(exDrugSyn, type = "surface", what = "LoeweExcess")

Dose-effect based synergy evaluation

$a/A+b/B=1$

$CI_{ab}=a/A+b/B$

Isobologram

isobologram(exDrugSyn, effect = 0.5)
isobologram(exDrugSyn, effect = 0.5, mode = "SP")

PieGes/synerdrug documentation built on May 8, 2019, 1:41 p.m.