highSCREEN (version 0.4)

rankhits: Rank hits

Description

This function ranks the candidate compound hits based on the mean of t1-specific scores.

Usage

rankhits(dat, score.before="S0", score.after="S1", var="m1")

Arguments

dat

Data frane as the output of the function \(formatRESULT\).

score.before

Specifies the name of the columns containing t0-specific scores.

score.after

Specifies the name of the columns containing t1-specific scores.

var

Specifies the variable according to which the ranking is done.

Value

Data frame with a row for each hit. The columns are as follows:

ID

Has the complete information to identify a compound or control. It contains information about the main plate, the quadrant/plate and the well.

MainPlate

Specifies the main plate to wich the compound/control belongs.

Plate

Specifies the quadrant/plate to wich the compound/control belongs.

Norm

Specifies the normalization method that was applied for the specific compound.

well

Specifies the location (row and column) of the compound/control in the quadrant.

row

Specifies the row location in the quadrant.

col

Specifies the column location in the quadrant.

welltype

Specifies if the well is compound or control.

S0

Replicates of the score from the t0-specific data set.

S1

Replicates of the score from the t1-specific data set.

IND2

Indicator variable specifying if the compound passes (\(TRUE\)) or fails (\(FALSE\)) QC2.

IND3

Indicator variable specifying if the compound passes (\(TRUE\)) or fails (\(FALSE\)) QC3.

diff

Difference between mean of t1 and t0 replicates.

m0

Mean of t0-specific replicates (MB).

s0

Standard deviation of t0-specific replicates (SB).

rs0

Coefficient of variation of t0-specific replicates. It is the ratio of S0 and M0.

m1

Mean of t1-specific replicates (MA).

s1

Standard deviation of t1-specific replicates (SA)

rs1

Coefficient of variation of t1-specific replicates. It is the ratio of S1 and M1.

ind_below

see \(ind\) below

ind_above

see \(ind\) below

ind

\(TRUE\) (if both \(ind_below\) and \(ind_above\) are \(TRUE\)) if RS1 is within the \(1.5 \times IQR\), where interquartile range (\(IQR\)) is computed based on all hits.

Examples

# NOT RUN {
set.seed(1234)
nc = 24
nr = 16

# create 1st replicate of data matrix with compounds and controls
r1 = matrix(abs(rnorm(nr*nc)*0.01), nr, nc)

# create control map
cmap = data.frame(X1=c(rep("Control P", floor(nr/3)), 
rep(c("Control low", "Control med", "Control high"), 
(floor(nr/3)+nr-3*floor(nr/3))/3), rep("Control N", 
floor(nr/3))), X2=c(rep("Control N", floor(nr/3)), 
rep(c("Control low", "Control med", "Control high"), 
(floor(nr/3)+nr-3*floor(nr/3))/3), rep("Control P", floor(nr/3))))
cmap = cmap[seq(1,nr,2),]

# create 2nd replicate of data matrix with compounds and controls
r2 = matrix(abs(rnorm(nr*nc)*0.01), nr, nc)

# create 3rd replicate of data matrix with compounds and controls
r3 = matrix(abs(rnorm(nr*nc)*0.01), nr, nc)

# combine all replicates for the t0-specific data
replicates_t0 = list(r1, r2, r3)
names(replicates_t0) = c("R1", "R2", "R3")

# create 1st replicate of data matrix with compounds and controls
r1 = matrix(abs(rnorm(nr*nc)), nr, nc)

# create 2nd replicate of data matrix with compounds and controls
r2 = matrix(abs(rnorm(nr*nc)), nr, nc)

# create 3rd replicate of data matrix with compounds and controls
r3 = matrix(abs(rnorm(nr*nc)), nr, nc)

# combine all replicates for the t1-specific data
replicates_t1 = list(r1, r2, r3)
names(replicates_t1) = c("R1", "R2", "R3")

# extract plate 1, replicate 1
dat1 = extractplate(replicates_t0, replicates_t1, plate=1, replicate=1)
# extract plate 1, replicate 2
dat2 = extractplate(replicates_t0, replicates_t1, plate=1, replicate=2)
# extract plate 1, replicate 3
dat3 = extractplate(replicates_t0, replicates_t1, plate=1, replicate=3)

# no normalizion
datraw1 = normplate("Main Plate 1", dat1[["dat0"]], dat1[["dat1"]], cmap,
 plate=1, replicate=1, norm="raw")
datraw2 = normplate("Main Plate 1", dat2[["dat0"]], dat2[["dat1"]], cmap,
 plate=1, replicate=2, norm="raw")
datraw3 = normplate("Main Pltae 1", dat3[["dat0"]], dat3[["dat1"]], cmap,
 plate=1, replicate=3, norm="raw")

# combine 3 replicates
datraw = rbind(datraw1, datraw2, datraw3)

# reformat result
datraw = formatRESULT(datraw, replicate="Replicate", t="Time")

# c-score normalization
datnorm1 = normplate("Main Plate 1", dat1[["dat0"]], dat1[["dat1"]], cmap,
 plate=1, replicate=1, norm="cscore", 
poscont="Control P", negcont="Control N")
datnorm2 = normplate("Main Plate 1", dat2[["dat0"]], dat2[["dat1"]], cmap,
 plate=1, replicate=2, norm="cscore", 
poscont="Control P", negcont="Control N")
datnorm3 = normplate("Main Pltae 1", dat3[["dat0"]], dat3[["dat1"]], cmap,
 plate=1, replicate=3, norm="cscore", 
poscont="Control P", negcont="Control N")

# combine 3 replicates
datnorm = rbind(datnorm1, datnorm2, datnorm3)

# reformat result
datnorm = formatRESULT(datnorm, replicate="Replicate", t="Time")

# identify hits
h = hits(datraw, datnorm, qc.mainplates="Main Plate 1", qc1.val=0.225, hit.val=3)

# rank hits in disending order of mean of t1-specific replicate scores "m1"
head(rankhits(h))
# }