Skip to content

'Record' row values every time the binary value in a collumn changes

5 messages · baboon2010, jim holtman, William Dunlap +2 more

#
My question is twofold.

Part 1:
My data looks like this:

(example set, real data has 2*10^6 rows)
binary<-c(1,1,1,0,0,0,1,1,1,0,0)
Chromosome<-c(1,1,1,1,1,1,2,2,2,2,2)
start<-c(12,17,18,20,25,36,12,15,16,17,19)
Table<-cbind(Chromosome,start,binary)
      Chromosome start binary
 [1,]          1    12      1
 [2,]          1    17      1
 [3,]          1    18      1
 [4,]          1    20      0
 [5,]          1    25      0
 [6,]          1    36      0
 [7,]          2    12      1
 [8,]          2    15      1
 [9,]          2    16      1
[10,]          2    17      0
[11,]          2    19      0

As output I need a shortlist for each binary block: giving me the starting
and ending position of each block.
Which for these example would look like this:
     Chromosome2 position_start position_end binary2
[1,]           1             12           18       1
[2,]           1             20           36       0
[3,]           2             12           16       1
[4,]           2             17           19       0

Part 2:
Based on the output of part 1, I need to assign the binary to rows of
another data set. If the position value in this second data set falls in one
of the blocks defined in the shortlist made in part1,the binary value of the
shortlist should be assigned to an extra column for this row.  This would
look something like this:
     Chromosome3 position Value binary3
 [1,] "1"         "12"     "a"   "1"    
 [2,] "1"         "13"     "b"   "1"    
 [3,] "1"         "14"     "c"   "1"    
 [4,] "1"         "15"     "d"   "1"    
 [5,] "1"         "16"     "e"   "1"    
 [6,] "1"         "18"     "f"   "1"    
 [7,] "1"         "20"     "g"   "0"    
 [8,] "1"         "21"     "h"   "0"    
 [9,] "1"         "22"     "i"   "0"    
[10,] "1"         "23"     "j"   "0"    
[11,] "1"         "25"     "k"   "0"    
[12,] "1"         "35"     "l"   "0"    
[13,] "2"         "12"     "m"   "1"    
[14,] "2"         "13"     "n"   "1"    
[15,] "2"         "14"     "o"   "1"    
[16,] "2"         "15"     "p"   "1"    
[17,] "2"         "16"     "q"   "1"    
[18,] "2"         "17"     "s"   "0"    
[19,] "2"         "18"     "d"   "0"    
[20,] "2"         "19"     "f"   "0"    


Many thanks in advance,

Niels

--
View this message in context: http://r.789695.n4.nabble.com/Record-row-values-every-time-the-binary-value-in-a-collumn-changes-tp3462496p3462496.html
Sent from the R help mailing list archive at Nabble.com.
#
Here is an answer to part 1:
+                      , list(Table[, "Chromosome"], Table[, 'binary'])
+                      , drop = TRUE
+                      )
+                   , function(.indx){
+     se <- range(.indx)
+     c(Chromosome2 = unname(Table[se[1L], "Chromosome"])
+       , position_start = unname(Table[se[1L], 'start'])
+       , position_end = unname(Table[se[2L], 'start'])
+       , binary2 = unname(Table[se[1L], 'binary'])
+       )
+ })
Chromosome2 position_start position_end binary2
1.0           1             20           36       0
2.0           2             17           19       0
1.1           1             12           18       1
2.1           2             12           16       1

        
On Wed, Apr 20, 2011 at 5:01 AM, baboon2010 <nielsvanderaa at live.be> wrote:

  
    
#
The following will likely be quicker way to find where
a column changes values than that lapply() when there
are lots of rows:

  f1 <- function (Table) {
      isFirstInRun <- function(x) c(TRUE, x[-1] != x[-length(x)])
      isLastInRun <- function(x) c(x[-1] != x[-length(x)], TRUE)
      with(data.frame(Table), {
          first <- isFirstInRun(binary)
          last <- isLastInRun(binary)
          cbind(Chromosome2 = Chromosome[first], position_start = start[first], 
              position_end = start[last], binary2 = binary[first])
    })
  }

E.g.,

  > f1(Table)
       Chromosome2 position_start position_end binary2
  [1,]           1             12           18       1
  [2,]           1             20           36       0
  [3,]           2             12           16       1
  [4,]           2             17           19       0

Bill Dunlap
Spotfire, TIBCO Software
wdunlap tibco.com
-binary-value-in-a-collumn-changes-tp3462496p3462496.html
#
Hi:

Here are a couple more options using packages plyr and data.table. The
labels in the second part are changed because they didn't make sense
in a 2M line file (well, mine may not either, but it's a start). You
can always change them to something more pertinent.

# Question 1:
Table <- data.frame(binary, chromosome = Chromosome, start)

library(plyr)
(df <- ddply(Table, .(chromosome, binary), summarise, position_start =
min(start),
         position_end = max(start)))
  chromosome binary position_start position_end
1          1      0             20           36
2          1      1             12           18
3          2      0             17           19
4          2      1             12           16

library(data.table)
dTable <- data.table(Table, key = 'chromosome, binary')
(dt <- dTable[, list(position_start = min(start),
               position_end = max(start)), by = 'chromosome, binary'])
     chromosome binary position_start position_end
[1,]          1      0             20           36
[2,]          1      1             12           18
[3,]          2      0             17           19
[4,]          2      1             12           16

## Question 2:

For plyr, it's easy to write a function that takes a generic input data frame
(in this case, a single line) and then outputs a data frame with
positions and labels.

tfun <- function(df) {
     diff <- with(df, position_end - position_start + 1)
     position <- with(df, seq(position_start, position_end))
     value <- paste(df$chromosome, df$binary, letters[1:diff], sep = '.')
     data.frame(chromosome = df$chromosome, position, value, binary = df$binary)
    }

# Then:
chromosome position value binary
1           1       20 1.0.a      0
2           1       21 1.0.b      0
3           1       22 1.0.c      0
4           1       23 1.0.d      0
5           1       24 1.0.e      0
6           1       25 1.0.f      0
7           1       26 1.0.g      0
8           1       27 1.0.h      0
9           1       28 1.0.i      0
10          1       29 1.0.j      0
11          1       30 1.0.k      0
12          1       31 1.0.l      0
13          1       32 1.0.m      0
14          1       33 1.0.n      0
15          1       34 1.0.o      0
16          1       35 1.0.p      0
17          1       36 1.0.q      0
18          1       12 1.1.a      1
19          1       13 1.1.b      1
20          1       14 1.1.c      1
21          1       15 1.1.d      1
22          1       16 1.1.e      1
23          1       17 1.1.f      1
24          1       18 1.1.g      1
25          2       17 2.0.a      0
26          2       18 2.0.b      0
27          2       19 2.0.c      0
28          2       12 2.1.a      1
29          2       13 2.1.b      1
30          2       14 2.1.c      1
31          2       15 2.1.d      1
32          2       16 2.1.e      1

# For data.table, one can apply the internals of tfun directly:

dt[, list(chromosome = chromosome, position = seq(position_start, position_end),
            value = paste(chromosome, binary,
                      letters[1:(position_end - position_start + 1)],
sep = '.'),
            binary = binary), by = 'chromosome, binary']
   chromosome binary chromosome.1 position value binary.1
            1      0            1       20 1.0.a        0
            1      0            1       21 1.0.b        0
            1      0            1       22 1.0.c        0
            1      0            1       23 1.0.d        0
            1      0            1       24 1.0.e        0
            1      0            1       25 1.0.f        0
            1      0            1       26 1.0.g        0
            1      0            1       27 1.0.h        0
            1      0            1       28 1.0.i        0
            1      0            1       29 1.0.j        0
            1      0            1       30 1.0.k        0
            1      0            1       31 1.0.l        0
            1      0            1       32 1.0.m        0
            1      0            1       33 1.0.n        0
            1      0            1       34 1.0.o        0
            1      0            1       35 1.0.p        0
            1      0            1       36 1.0.q        0
            1      1            1       12 1.1.a        1
            1      1            1       13 1.1.b        1
            1      1            1       14 1.1.c        1
            1      1            1       15 1.1.d        1
            1      1            1       16 1.1.e        1
            1      1            1       17 1.1.f        1
            1      1            1       18 1.1.g        1
            2      0            2       17 2.0.a        0
            2      0            2       18 2.0.b        0
            2      0            2       19 2.0.c        0
            2      1            2       12 2.1.a        1
            2      1            2       13 2.1.b        1
            2      1            2       14 2.1.c        1
            2      1            2       15 2.1.d        1
            2      1            2       16 2.1.e        1
cn chromosome binary   chromosome position value   binary

HTH,
Dennis
On Wed, Apr 20, 2011 at 2:01 AM, baboon2010 <nielsvanderaa at live.be> wrote:
#
Here's one way to do part 1:
[,1] [,2] [,3] [,4]
[1,]    1   12   18    1
[2,]    1   20   36    0
[3,]    2   12   16    1
[4,]    2   17   19    0

If I understand you correctly, here's a way to do part 2:
[1] 1 1 1 1 1 1 0 0 0 0 0 0 1 1 1 1 1 0 0 0

 					- Phil Spector
 					 Statistical Computing Facility
 					 Department of Statistics
 					 UC Berkeley
 					 spector at stat.berkeley.edu