rankhits: Rank hits

In highSCREEN: High-Throughput Screening for Plate Based Essays

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

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))

highSCREEN documentation built on Feb. 12, 2021, 5:09 p.m.