sdRain: Add rain to a classification by using a chosen multiple of...

In CRUKMI-ComputationalBiology/twoddpcr: Classify 2-d Droplet Digital PCR (ddPCR) data and quantify the number of starting molecules

Description

Although we can use various algorithms to classify all droplets in a ddPCR experiment, there will be some variation between the classifications. We can perhaps have a relatively high confidence that droplets near the centres of clusters do indeed belong to that cluster, whereas we probably have a lower confidence in the classification of those further away, say, near the 'boundary' of two clusters. We may view these droplets (or a subset of them) as having an ambiguous class. This function allows us to only consider droplets classified within a certain distance of the means of each cluster and label the rest as "Rain".

Usage

sdRain(droplets, cMethod, errorLevel = 5, ...)

## S4 method for signature 'data.frame'
sdRain(droplets, cMethod, errorLevel = 5, fullTable = TRUE)

## S4 method for signature 'ddpcrWell'
sdRain(droplets, cMethod, errorLevel = 5)

## S4 method for signature 'ddpcrPlate'
sdRain(droplets, cMethod, errorLevel = 5)

Arguments

droplets	A ddpcrWell or ddpcrPlate object, or a droplet data frame including a classification column.
cMethod	The name or column number of the classification for which we want to add rain to.
errorLevel	How many multiples of standard deviation from the mean of each cluster to retain. Can be a list where each item corresponds to a class name and the multiple for that class. Can also be a numeric vector of length 1, which is equivalent to a list with all the same entries. Defaults to 5.
...	Other options depending on the type of droplets.
fullTable	If TRUE, returns a full data frame of droplets with an extra column of rainy data; if FALSE, simply returns a factor where each entry corresponds to an entry in the original classification column with added rain. Defaults to FALSE.

Value

If droplets is a data frame, return a data frame or factor (depending on fullTable) where droplets with ambiguous classifications are labelled as "Rain".

If droplets is a ddpcrWell object, return a ddpcrWell object with a rainy classification.

If droplets is a ddpcrPlate object, return a ddpcrPlate object with a rainy classifications.

Author(s)

Anthony Chiu, [e-mail]

References

This approach was described in Jones, M., Williams, J., Gaertner, K., Phillips, R., Hurst, J., & Frater, J. (2014). Low copy target detection by Droplet Digital PCR through application of a novel open access bioinformatic pipeline, "definetherain." Journal of Virological Methods, 202(100), 46–53. http://doi.org/10.1016/j.jviromet.2014.02.020

Examples

## Compare the types of droplets in a single well for the "Cluster" class
## and then with rain.
aWell <- ddpcrWell(well=KRASdata[["E03"]])
aWell <- sdRain(aWell, cMethod="Cluster")
cl <- wellClassification(aWell)
table(cl$Cluster)
table(cl$ClusterSdRain)

## Compare the types of droplets in multiple wells for the "Cluster" class
## and then with rain.
krasPlate <- ddpcrPlate(wells=KRASdata[c("E03", "H03", "C04", "F04")])
krasPlate <- sdRain(krasPlate, cMethod="Cluster")
plateSummary(krasPlate, cMethod="Cluster")[, c(1:5)]
plateSummary(krasPlate, cMethod="ClusterSdRain")[, c(1:5)]

## The 'errorLevel' parameter can changed.
krasPlate <- sdRain(krasPlate, cMethod="Cluster", errorLevel=4)
plateSummary(krasPlate, cMethod="ClusterSdRain")[, c(1:5)]

## The 'errorLevel' parameter can also be changed for each cluster.
krasPlate <- sdRain(krasPlate, cMethod="Cluster",
                    errorLevel=list(NN=5, NP=5, PN=4, PP=3))
plateSummary(krasPlate, cMethod="ClusterSdRain")[, c(1:5)]

CRUKMI-ComputationalBiology/twoddpcr documentation built on Feb. 14, 2021, 9:18 p.m.