In anwarbio/bioassays: Summarising Multi Well Plate Cellular Assay

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(bioassays)

Introduction

In a cell culture lab various cellular assays are performed. The package "bioassays" will help to analyse the results of these experiments performed in multiwell plates. The usage of various functions in the "bioassays" package is provided in this article.

The functions in this package can be used to summarise data from any multiwell plate, and by incorporating them in a loop several plates can be analyzed automatically. Two examples are also provided in the article.

prerequisite

The output reading from the instrument (eg.spectrophotometer) should be in a matrix format. An example data file (csv format) is shown below. If the data is in .xls/.xlsx format read_excel function from the readxl package can be used.

data(rawdata96)
head(rawdata96)

A metadata file is needed for the whole experiment. The row and col columns must be present in the metadata file to indicate the location of well. An example is given below.

data(metafile96)
head(metafile96)

Functions

1. Function: extract_filename

extract_filename helps to extract information from the file name. syntax is extract_filename(filename,split = " ",end = ".csv", remove = " ", sep="-"). filename is the file name. split is the portions at which the name has to be split (default is space " "). end is the extension of file name that need to be removed (default is ".csv"). remove is the portion from the file name that need to be omitted after splitting (default is space " "). sep add a symbol between separate sections, default is "-".

This function is useful for extracting specific information from file names, like compound name, plate number etc, to provide appropriate analysis.

For e.g.

extract_filename("L HEPG2 P3 72HRS.csv")

extract_filename("L HEPG2 P3 72HRS.csv", split=" ",end=".csv",remove="L",sep="")

2. Function: rmodd_summary

rmodd_summary helps to remove outliers and summarises the values from a given set of values. The syntax is rmodd_summary(x, rm = "FALSE", strict= "FALSE", cutoff=80,n=3). x is a numeric vector. rm = TRUE if want to remove outliers. If strict = FALSE values above/below 1.5 IQR are omitted (outliers omitted). If strict = TRUE more aggresive outlier removal is used to bring the %cv below the cutoff. n is the minimum number of samples you need per group if more aggressive outlier removal is used.

For example:

x<- c(1.01,0.98,0.6,0.54,0.6,0.6,0.4,3)
rmodd_summary(x, rm = "FALSE", strict= "FALSE", cutoff=80,n=3)

rmodd_summary(x, rm = "TRUE", strict= "FALSE", cutoff=80,n=3)

rmodd_summary(x, rm = "TRUE", strict= "TRUE", cutoff=20,n=5)

3. Function: data2plateformat

data2plateformat converts the data (eg: readings from a 96 well plate) to an appropriate matrix format. Syntax is data2plateformat(data, platetype = 96). data is the data to be formatted. platetype is the plate from which the data is coming. It can take 6, 12, 24, 96 or 384 values to represent the corresponding multiwell plate type.

For e.g. To rename columns and rows of rawdata96 to right format.

rawdata<-data2plateformat(rawdata96,platetype = 96)
head(rawdata)

4. Function: plate2df

plate2df coereces a matrix with values from multiwell plates into data.frame. The function uses column names and row names of datamatrix (2D data of a mutliwell plate) and generates a dataframe with row, col (column) and position indices. The value column represents the corresponding value in the datamarix.

Syntax is plate2df(datamatrix). datamatrix is the data in matrix format.

For eg.

OD_df <- plate2df(rawdata)
head(OD_df)

5. Function: matrix96

matrix96 helps to convert a data.frame into a matrix. The syntax is matrix96(dataframe,column,rm="FALSE"). dataframe is the data.frame to be formatted. The data.frame must contain row and col columns. column is the name of column that needs to be converted into a matrix. If rm= "TRUE" then -ve and NA are set to 0.

For e.g.

matrix96(OD_df,"value")

matrix96(OD_df,"position")

6. Function: plate_metadata

plate_metadata combines the plate-specific information (like compound used, standard concentration, dilution of samples, etc) and metadata into unique plate metadata. The syntax is plate_metadata(plate_details, metadata,mergeby="type"). plate details is the plate specific information that needs to be added to metadata. metadata is the metadata for the whole experiment. mergeby is the column that is common to both metadata and plate_meta (this column will be used to merge the information).

For eg. An incomplete meta data

head(metafile96)

Plate specific details are.

plate_details <- list("compound" = "Taxol",
                "concentration" = c(0.00,0.01,0.02,0.05,0.10,1.00,5.00,10.00),
                "type" = c("S1","S2","S3","S4","S5","S6","S7","S8"),
                "dilution" = 1)

Using the plate-specific info, the metadata can be filled by calling plate_metadata function.

plate_meta<-plate_metadata(plate_details,metafile96,mergeby="type")
head(plate_meta)

To join both plate_meta and OD_df, the dplyr::inner_join function can be used.

data_DF<- dplyr::inner_join(OD_df,plate_meta,by=c("row","col","position"))

head(data_DF)

7. Function: heatplate

heatplate creates a heatmap representation of a multiwell plate. The syntax is heatplate(datamatrix,name,size=7.5). datamatrix is the data in matrix format. An easy way to create this is by calling the matrix96 function, as explained before. name is the name to be given for heatmap, size is the size of each well in the heatmap (default is 7.5).

This function will produce a heatmap of normalized values if the variable is numeric. If it is a factor, it will simple provide a coloured categorical plot.

eg 1. Categorical plot

datamatrix<-matrix96(metafile96,"id")
datamatrix

heatplate(datamatrix,"Plate 1", size=5)

eg 2. Heatmap

rawdata<-data2plateformat(rawdata96,platetype = 96)
OD_df<- plate2df(rawdata)
data<-matrix96(OD_df,"value")
data

heatplate(data,"Plate 1", size=5)

8. Function: reduceblank

reduceblank helps to reduce blank values from the readings.

The syntax is reduceblank (dataframe,x_vector,blank_vector,y). dataframe is the data. x_vector are the entries for which the blank has to be subtracted If all entries have to subtracted, then use "All". x_vector should be a vector eg: c("drug1","drug2",drug3" etc). blank_vector is a vector of blank names whose value has to be subtracted eg: c("blank1","blank2","blank3","blank4")). This function will subtract the first blank vector element from first x_vector element and so on. y is the column name where the action will take place. y should be numeric. The results will appear as a new column named blankminus.

For eg.

data_DF<-reduceblank(data_DF, x_vector =c("All"),
                     blank_vector = c("Blank"), "value")
head(data_DF)

9. Function: estimate

estimate estimates an unknown variable (eg: concentration) based on the standard curve. Syntax is estimate(data=dataframe,colname="blankminus",fitformula=fit, methord="linear/nplr"). data is the dataframe which needs to be evaluated. colname is the column name for which the values has to be estimated. fitformula represents the model to be fit methord specifies if linear or nonparametric logistic curve was used for the fitformula.

For eg: data_DF is a dataframe for which the concentration has to be estimated based on the value of blankminus.

For filtering the standards

std<- dplyr::filter(data_DF, data_DF$id=="STD")  
std<- aggregate(std$blankminus ~ std$concentration, FUN = mean )
colnames (std) <-c("con", "OD")
head(std)

To fit a standard curve:

fit1 is the 3 parameter logistic curve model and fit2 is the linear regression model. The appropriate one for your experiment can be used.

fit2<-stats::lm(formula = con ~ OD,data = std)# linear model
fit1<-nplr::nplr(std$con,std$OD,npars=3,useLog = FALSE)#  nplr, 3 parameter model

For estimating the concentration using linear model

estimated<-estimate(data_DF,colname="blankminus",fitformula=fit2,method="linear")
head(estimated)

For estimating the concentration using nplr methord

estimated2<-estimate(data_DF,colname="blankminus",fitformula=fit1,method="nplr")
head(estimated2)

10. Function: dfsummary

dfsummary() summarizes the data.frame (based on a column). It has additional controls to group samples and to omit variables not needed. The syntax is dfsummary(dataframe,y,grp_vector,rm_vector,nickname,rm="FALSE",param). dataframe is the data. y is the numeric variable (column name) that has to be summarized. grp_vector is a vector of column names, determining which samples will be grouped. The order of the elements in grp_vector determines the order of grouping. rm_vector is the vector of items to be omitted before summarizing. nickname is the name for the output data.frame. Set rm="FALSE" if outliers should not be removed. To apply more stringent thresholds for removing outliers, parameters can be provided in the param vector. param has to be entered in the format c(strict="TRUE",cutoff=40,n=12). For details please refer to the rmodd_summaryfunction.

In the following example, the data has to be summarized based on the "type" column. "estimated" values are summarized. Samples are grouped as per "id". "STD" and "Blank" values need to be omitted. Outliers are retained (rm="FALSE"). The nickname for the plate is "plate1".

result<-dfsummary(estimated,"estimated",c("id","type"),
        c("STD","Blank"),"plate1", rm="FALSE",
        param=c(strict="FALSE",cutoff=40,n=12))

result

11. Function: pvalue

The pvalue() function applies a t-test to the result dataframe. The syntax is pvalue(dataframe,control,sigval). dataframe is the result of dfsummary. control is the group that is considered as control, sigval is the pvalue cutoff (a value below this is considered as significant).

For example:

pval<-pvalue(result, control="S8", sigval=0.05)
head(pval)

anwarbio/bioassays documentation built on Oct. 9, 2020, 10:35 p.m.