molecular biology

QPix2 Part 2

The QPix software identifies a single region per plate (two regions total) and randomly picks a user defined number of colonies from each region. Ideally one would want to define 96 regions per plate so that the number of colonies per region can be specified - not allowed. However each colony’s location (X|Y deck coordinate) is recorded in an XML log file.

With a bit of hacking, the well location can be assigned to each colony. In this post I will process the XML log file to extract X|Y coordinates for each picked colony, associate those coordinates with wells, associate clone IDs with wells and ultimately associate the clone ID with a picked colony.

See this post for details concerning the input files and custom methods.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
rm(list=ls(all=TRUE))
library(XML)

setwd('~./qpix2')
#this is the files that indicates the clone IDs of the VH, VL fragments that
#have been cloned - the ID of the transformation mix
assoc.file <- "amplicon-plate-map.txt"
assoc <- read.table( paste(getwd(),"input", assoc.file, sep="/"), skip=0, header=TRUE)

 > assoc
         id dest.plate dest.well
1  PL101-VH  amplicons       A01
2  PL102-VH  amplicons       B01
3  PL103-VH  amplicons       C01
4   PL18-VH  amplicons       D01
5   PL28-VH  amplicons       E01
6   PL38-VH  amplicons       F01


##extract the runid as six integers
qpix.filename <-  "QSoft Log 2014-02-11 105201.XML"
runid <- substring( qpix.filename, 22, 27) 
> runid
[1] "105201"

Supply plate names of the plates in which the picked colonies were deposited. This is not needed if the plate names are typed directly into the software. For this example I will assume they are not. Enter as many destination plates as exist, the script will adjust

1
2
3
4

dest.plate <- list()
dest.plate[[1]] <- "PL140211c"
dest.plate[[2]] <- "PL140211d"

Use region 2 well H01 as the “origin”. Enter calibration block values for well G01. From G01 calculate H01 using hard coded offsets. 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

g01.x <- 114278
g01.y <- 241732
xstart <- g01.x + 2580
ystart <- g01.y - 8781
> xstart
[1] 116858
> ystart
[1] 232951

d <- xmlTreeParse( paste(getwd(), "/qpix-logs/", qpix.filename, sep="") )
top <- xmlRoot(d)
##top[[11]] has the "contents"; should be as many deposits as destination plates
>names(top[[11]]) 
     routine      imaging   colonydata      deposit      deposit 
   "routine"    "imaging" "colonydata"    "deposit"    "deposit" 
>

Retrieve XML log files off the Qpix post run. Colony data is in //contents/colonydata. The XML can be easily visulaized by opening with a browser. Colonydata contains X|Y location as well as region. Read it in and create a dataframe. 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
colony.data <- getNodeSet(top, "//contents/colonydata")
d2 <- data.frame(t(sapply( xmlApply(colony.data[[1]],  xmlAttrs),c)), stringsAsFactors=FALSE)
# region-image-colonyid uniquely identifies a colony
# not sure why image is needed
d2 <- d2[,c("colonyid","image","posx","posy","region")]

 > head(d2)
  colonyid image   posx   posy region
1        0     5 159286 385817      1
2        1     5 159569 403677      1
3        2     5 160654 385588      1
4        3     5 160761 388800      1
5        4     5 160854 371823      1
6        5     5 161284 390663      1
>

Calculate centers of each column along X axis.
Offset between regions 1 and 2 is 108174um.
The custom method getDeckWell defined here
Add the source well to the dataframe.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37

x.cent <- seq( xstart, by=9100, length.out=12)
yr1.cent <- seq( ystart+108174, by=9100, length.out=8)
yr2.cent <- seq( ystart, by=9100, length.out=8)

src.well <- rep( NA, nrow(d2))

for( i in 1:nrow(d2)){
  src.well[i] <- getDeckWell( as.numeric(d2[ i, "posx"][[1]]), as.numeric(d2[ i, "posy"][[1]]), as.numeric(d2[ i, "region"][[1]])) 
}
d2 <- cbind( d2, src.well)

> head(d2)
  colonyid image   posx   posy region src.well
1        0     5 159286 385817      1      C06
2        1     5 159569 403677      1      A06
3        2     5 160654 385588      1      C06
4        3     5 160761 388800      1      C06
5        4     5 160854 371823      1      E06
6        5     5 161284 390663      1      C06
> 

d2.wid <- merge( d2, assoc, by.x="src.well", by.y="dest.well")

> d2.wid
          id colonyid src.well
1   PL101-VH      153      A01
2   PL101-VH      162      A01
3   PL101-VH      157      A01
4   PL101-VH      166      A01
5   PL101-VH      160      A01
6    PL44-VH      208      A02
7    PL44-VH      181      A02
8    PL44-VH      206      A02
9    PL44-VH      187      A02
10   PL44-VH      171      A02

Now we have associated an id based on well location to each colony.
src.well here is the well the clone came from in the transformation plate.
All we need here is colonyid assigned by Qpix and my clone id
src.well is the transformation well, keep that so you can
retrieve colonies that weren’t recovered by the Qpix.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54

d2.wid <- d2.wid[,c("id","colonyid","src.well")]
deposit <-getNodeSet(top, "//contents/deposit")

deposit.list <- list()
for( i in 1: length(deposit)){
  deposit.list[[i]] <- data.frame(t(sapply( xmlApply(deposit[[i]],  xmlAttrs),c)), stringsAsFactors=FALSE)
}

#may not need to do this if destinations entered in software
for(i in 1:length(deposit.list)){
  deposit.list[[i]]$destination <- dest.plate[[i]]
}

d3 <- do.call("rbind", deposit.list)


d3$location <- sapply(  d3$location, insert0inWell)

> head(d3)
  destination location colonyid              source region
1   PL140211c      A01        0 UIDUO0211-3912159-X      1
2   PL140211c      B01        1 UIDUO0211-3912159-X      1
3   PL140211c      C01        2 UIDUO0211-3912159-X      1
4   PL140211c      D01        3 UIDUO0211-3912159-X      1
5   PL140211c      E01        4 UIDUO0211-3912159-X      1
6   PL140211c      F01        5 UIDUO0211-3912159-X      1
 
d4 <- merge(d3, d2.wid, by.x="colonyid", by.y="colonyid")
d4 <- d4[,c("destination","location","id","src.well")] #src.well is the transformation well 
d4 <- d4[ order( d4$destination, d4$location),]
names(d4)[1:2] <- c("plate","well")
> head(d4)
        plate well      id src.well
1   PL140211c  A01 PL78-VL      C06
137 PL140211c  A02 PL44-VL      A06
51  PL140211c  A03 PL86-VL      C07
86  PL140211c  A04 PL86-VL      C07
95  PL140211c  A05 PL95-VL      D08
104 PL140211c  A06 PL94-VL      C08

results <- as.data.frame(table(d4$src.well))
results <- results[ order( results$Freq),]

> results
   Var1 Freq
1   A01    0
15  C02    0
22  C09    0
26  D04    0
31  D09    0
40  E09    0
41  E10    0

Results shows the count of colonies obtained per clone ID. Some are not present in the collection, as can be seen above where Freq==0. This provides a list of wells that can be plated manually if these are considered valuable colonies.

Pick out 4 candidates (if available) for each clone ID.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
d5 <- do.call("rbind", lapply( split( d4, d4$bdn ), function(x) x[1:4,]) )
d5 <- d5[ !is.na(d5$bdn),]
d5 <- d5[ order( d5$plate, d5$well ),]

##evaluate
set1 <- read.table( paste(getwd(), "103410-picked-plates-map.txt", sep=""), skip=0, header=TRUE)
set2 <- read.table( paste(getwd(), "105201-picked-plates-map.txt", sep=""), skip=0, header=TRUE)

composite <- rbind(set1, set2)
out.file <-paste( working.dir,  "composite.txt", sep="")
write.table( composite, file = out.file, append = FALSE, quote = FALSE, sep = "\t", eol = "\n", na = "NA", dec = ".", row.names = FALSE, col.names = TRUE, qmethod = c("escape", "double"))



results <- as.data.frame(table(composite$src.well))
results <- results[ order( results$Freq),]

plot(results)

set1 and set2 provide the plate maps for the picked plates.
results show the number of colonies picked for each candidate id. For those with 0-2 colonies, manually pick supplemental colonies if available.

results can be plotted to get a sense for the distribution of colony counts for the various candidate clones.

Share