CODEWITHSUNDEEP
rlapply
fugu
fugu

Reputation: 6568

lapply on several subsets of data frame

I have a dataframe data containing the chromosome and position of mutated nucleotides within a genome:

structure(list(chrom = c(1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 3L, 
3L, 4L, 4L, 4L, 4L), pos = c(10L, 200L, 134L, 400L, 600L, 1000L, 
20L, 33L, 40L, 45L, 50L, 55L, 100L, 123L)), .Names = c("chrom", 
"pos"), class = "data.frame", row.names = c(NA, -14L))

  chrom  pos
1     1   10
2     1  200
3     1  134
4     1  400
5     1  600
6     1 1000

And another tss_locations, containing the location of a feature (tss) within a gene and the chromosome it's on:

structure(list(gene = structure(c(1L, 4L, 5L, 6L, 7L, 8L, 9L, 
10L, 11L, 2L, 3L), .Label = c("gene1", "gene10", "gene11", "gene2", 
"gene3", "gene4", "gene5", "gene6", "gene7", "gene8", "gene9"
), class = "factor"), chrom = c(1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 
3L, 4L, 4L), tss = c(5L, 10L, 23L, 1340L, 313L, 88L, 44L, 57L, 
88L, 74L, 127L)), .Names = c("gene", "chrom", "tss"), class = "data.frame", row.names = c(NA, 
-11L))

   gene chrom  tss
1 gene1     1    5
2 gene2     1   10
3 gene3     1   23
4 gene4     2 1340
5 gene5     2  313
6 gene6     2   88

I'm trying to calculate the distance to the closest tss on the same chromosome for each pos in data.

So far, I can calculate the distance from each data$pos to any tss_locations$tss (i.e. the closest tss to each pos irrespective of chromosome): 

fun <- function(p) {
  # Get index of nearest tss
  index<-which.min(abs(tss_locations$tss - p))
  # Lookup the value
  closestTss<-tss_locations$tss[[index]]
  # Calculate the distance
  dist<-(closestTss-p)
  list(snp=p, closest=closestTss, distance2nearest=dist)
}

# Run function for each 'pos' in data
dist2tss<-lapply(data$pos, fun)

# Convert to data frame and sort descending:
dist2tss<-do.call(rbind, dist2tss)
dist2tss<-as.data.frame(dist2tss)

dist2tss<-arrange(dist2tss,(as.numeric(distance2nearest)))
dist2tss$distance2nearest<-as.numeric(dist2tss$distance2nearest)

head(dist2tss)

  snp closest distance2nearest
1 600     313             -287
2 400     313              -87
3 200     127              -73
4 100      88              -12
5  33      23              -10
6 134     127               -7

However, I would like to be able to find the closest tss on the same chromosome for each pos.

I know that I could apply this to each chromosome individually, but I want to see the global distances (across all chromosomes), but only make comparisons between positions and tsss on the same chromosome.

How can I adjust this to achieve this goal? Subset both data frames by chromosome and merge the results?

Is this the right approach so far?

Upvotes: 0

Views: 82

Answers (1)

Matt Jewett
Matt Jewett

Reputation: 3369

Something like this may work to get the closest tss for each chromosome in the data dataframe.

data <- structure(list(chrom = c(1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 3L, 
                                 3L, 4L, 4L, 4L, 4L), pos = c(10L, 200L, 134L, 400L, 600L, 1000L, 
                                                              20L, 33L, 40L, 45L, 50L, 55L, 100L, 123L)), .Names = c("chrom", 
                                                                                                                     "pos"), class = "data.frame", row.names = c(NA, -14L))

tss_locations <- structure(list(gene = structure(c(1L, 4L, 5L, 6L, 7L, 8L, 9L, 
                                                  10L, 11L, 2L, 3L), .Label = c("gene1", "gene10", "gene11", "gene2", 
                                                                                "gene3", "gene4", "gene5", "gene6", "gene7", "gene8", "gene9"
                                                  ), class = "factor"), chrom = c(1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 
                                                                                  3L, 4L, 4L), tss = c(5L, 10L, 23L, 1340L, 313L, 88L, 44L, 57L, 
                                                                                                       88L, 74L, 127L)), .Names = c("gene", "chrom", "tss"), class = "data.frame", row.names = c(NA, 
                                                                                                                                                                                                 -11L))

# Generate needed values by applying function to all rows and transposing t() the results
data[,c("closest_gene", "closest_tss", "min_dist")] <- t(apply(data, 1, function(x){
   # Get subset of tss_locations where the chromosome matches the current row
   genes <- tss_locations[tss_locations$chrom == x["chrom"], ]

   # Find the minimum distance from the current row's pos to the nearest tss location
   min.dist <- min(abs(genes$tss - x["pos"]))

   # Find the closest tss location to the current row's pos
   closest_tss <- genes[which.min(abs(genes$tss - x["pos"])), "tss"]

   # Check if closest tss location is less than pos and set min.dist to negative if true
   min.dist <- ifelse(closest_tss < x["pos"], min.dist * -1, min.dist)

   # Find the closest gene to the current row's pos
   closest_gene <- as.character(genes[which.min(abs(genes$tss - x["pos"])), "gene"])

   # Return the values to the matrix
   return(c(closest_gene, closest_tss, min.dist))
}))

Upvotes: 1

Related Questions