map gene positions to chromosome coordinates

Question

First post here so I hope I can explain myself at the best.

I need to cross-reference two dataframes by finding if one specific chromosome location given in one of the two dataframes occurs in the range provided by the other one, and as result I would like to have a new column with the gene present in that range.

"genes"is the dataframe with the coordinates (start/end) to be considered as the range

head(genes)
# A tibble: 6 x 9
  chr   source         type      start       end strand gene_id         symbol        gene_biotype  
  <chr> <chr>          <chr>     <int>     <int> <chr>  <chr>           <chr>         <chr>         
1 2     pseudogene     gene  143300987 143301544 +      ENSG00000228134 AC092578.1    pseudogene    
2 2     pseudogene     gene  143611664 143613567 +      ENSG00000229781 AC013444.1    pseudogene    
3 2     protein_coding gene  143635067 143799890 +      ENSG00000115919 KYNU          protein_coding
4 2     pseudogene     gene  143704869 143705655 -      ENSG00000270390 RP11-470B22.1 pseudogene    
5 2     miRNA          gene  143763269 143763360 -      ENSG00000221169 AC013444.2    miRNA         
6 2     protein_coding gene  143848931 144525921 +      ENSG00000075884 ARHGAP15      protein_coding

the other data frame(x) is:

  chr_a   point A
1     2 143301002 
2     2 143625061
3     2 143700941
4     2 143811317
5     2 144127323
6     2 144224689

I basically have to find whether "point A" falls between "start"/ "end" range in (genes) and which gene symbol is associated.

I tried the following:

x$geneA <- ifelse(sapply(x$`point A`, function(g)
  any(genes$start >= g & genes$end <=g)), genes$symbol, NA)

but the results I get are not in line with the genomic coordinates.

Hope someone can help me! Thx!

Does this answer your question? [Efficiently merging two data frames on a non-trivial criteria](https://stackoverflow.com/questions/18840410/efficiently-merging-two-data-frames-on-a-non-trivial-criteria) and [Overlap join with start and end positions](https://stackoverflow.com/questions/24480031/overlap-join-with-start-and-end-positions) — A. Suliman, Feb 21 '20 at 15:14

Aron Strandberg · Accepted Answer · 2020-02-21T15:06:12.583

Does this work?

I'm assuming that each point matches to only one gene symbol.

x$geneA <- sapply(x$`point A`,
                  function(g) filter(genes, g >= start & g <= end)$symbol[1])

Result:

x

# A tibble: 6 x 3
  chr_a `point A` geneA     
  <int>     <int> <chr>     
1     2 143301002 AC092578.1
2     2 143625061 NA        
3     2 143700941 KYNU      
4     2 143811317 NA        
5     2 144127323 ARHGAP15  
6     2 144224689 ARHGAP15

score 2 · Answer 2 · answered Feb 21 '20 at 15:11

welcome to Stackoverflow! In the future, please post a minimal, workable example (MWE).

genes <- tribble(~chr, ~source, ~type, ~start, ~end, ~strand, ~gene_id, ~symbol, ~gene_biotype,
                  2, "pseudogene", "gene", 143300987, 143301544, "+", "ENSG00000228134", "AC092578.1", "pseudogene",
                  2, "pseudogene", "gene", 143611664, 143613567, "+", "ENSG00000229781", "AC013444.1", "pseudogene",
                  2, "protein_coding", "gene", 143635067, 143799890, "+", "ENSG00000115919", "KYNU", "protein_coding",
                  2, "pseudogene", "gene", 143704869, 143705655, "-", "ENSG00000270390", "RP11-470B22.1", "pseudogene",
                  2, "miRNA", "gene", 143763269, 143763360, "-", "ENSG00000221169", "AC013444.2", "miRNA",
                  2, "protein_coding", "gene", 143848931, 144525921, "+", "ENSG00000075884", "ARHGAP15", "protein_coding")

x <- tribble(~chr_a, ~`point A`,
              2, 143301002,
              2, 143625061,
              2, 143700941,
              2, 143811317,
              2, 144127323,
              2, 144224689,
)

I give you a tidyverse approach:

x %>% 
    nest_join(genes, by = c("chr_a" = "chr")) %>% 
    group_by(`point A`) %>% 
    mutate(genes = map(genes, ~filter(., `point A` >= start & `point A` <= end))) %>% 
    unnest(genes, keep_empty = TRUE)

for getting a merged table where the non-matching rows are NA. Or simply find the ones that do match without using nested tibbles

x %>% 
    left_join(genes, by = c("chr_a" = "chr")) %>% 
    filter(`point A` >= start & `point A` <= end)

score 1 · Answer 3 · answered Feb 21 '20 at 15:13

You can try the base R code below

df2out <- within(df2,symbol <- sapply(A, function(x) df1$symbol[which(x>=df1$start & x<=df1$end)]))

such that

> df2out
  chr_a point         A     symbol
1     1     2 143301002 AC092578.1
2     2     2 143625061           
3     3     2 143700941       KYNU
4     4     2 143811317           
5     5     2 144127323   ARHGAP15
6     6     2 144224689   ARHGAP15

DATA

df1 <- structure(list(chr = c(2L, 2L, 2L, 2L, 2L, 2L), source = c("pseudogene", 
"pseudogene", "protein_coding", "pseudogene", "miRNA", "protein_coding"
), type = c("gene", "gene", "gene", "gene", "gene", "gene"), 
    start = c(143300987L, 143611664L, 143635067L, 143704869L, 
    143763269L, 143848931L), end = c(143301544L, 143613567L, 
    143799890L, 143705655L, 143763360L, 144525921L), strand = c("+", 
    "+", "+", "-", "-", "+"), gene_id = c("ENSG00000228134", 
    "ENSG00000229781", "ENSG00000115919", "ENSG00000270390", 
    "ENSG00000221169", "ENSG00000075884"), symbol = c("AC092578.1", 
    "AC013444.1", "KYNU", "RP11-470B22.1", "AC013444.2", "ARHGAP15"
    ), gene_biotype = c("pseudogene", "pseudogene", "protein_coding", 
    "pseudogene", "miRNA", "protein_coding")), class = "data.frame", row.names = c(NA, 
-6L))

df2 <- structure(list(chr_a = 1:6, point = c(2L, 2L, 2L, 2L, 2L, 2L), 
    A = c(143301002L, 143625061L, 143700941L, 143811317L, 144127323L, 
    144224689L)), class = "data.frame", row.names = c(NA, -6L
))

score 1 · Answer 4 · answered Feb 21 '20 at 15:30

Most likely this answer will never be seen =p

There are packages for this. Note, your code will not work with extra chromosomes, or strands.

Using data by @koenniem,

library(GenomicRanges)

gr1 = makeGRangesFromDataFrame(genes,keep.extra.columns=TRUE)

x = data.frame(x,check.names=FALSE)
gr2 = GRanges(seqnames=x$chr_a,IRanges(start=x[,"point A"],width=1))

x$gene = NA
ovlp = findOverlaps(gr2,gr1)
x$gene[queryHits(ovlp)] = gr1$symbol[subjectHits(ovlp)]

  chr_a   point A       gene
1     2 143301002 AC092578.1
2     2 143625061       <NA>
3     2 143700941       KYNU
4     2 143811317       <NA>
5     2 144127323   ARHGAP15
6     2 144224689   ARHGAP15

score 0 · Answer 5 · answered Feb 21 '20 at 15:09

A for-loop based solution. (This is, of course, going to be much slower than using an apply.)

#A mock-up of your data
symbol <- c("AC092578.1", "AC013444.1", "KYNU", "RP11-470B22.1", "AC013444.2", "ARHGAP15", "Newadditionalsymbol")
start <- c(143300987, 143611664, 143635067, 143704869, 143763269, 143848931, 143300987)
end <- c(143301544, 143613567, 143799890, 143705655, 143763360, 144525921, 143301044)

genes <- data.frame(start, end, symbol, stringsAsFactors = F)

point_A <- start[1:6]+1
chr_1 <- rep_len(2, length.out = length(point_A))

x <- data.frame(chr_1, point_A, stringsAsFactors = F)

x$symbol <- NA #Create a new column to store the symbols, populate it with NA

x

#      chr_1   point_A symbol
# 1     2 143300988     NA
# 2     2 143611665     NA
# 3     2 143635068     NA
# 4     2 143704870     NA
# 5     2 143763270     NA
# 6     2 143848932     NA

#Solution using a for loop
for(i in 1:nrow(x)){ #Iterate through every row of x

  for(j in 1:nrow(genes)){ #Iterate through every row of genes

    if(x$point_A[i] >= genes$start[j] & x$point_A[i] < genes$end[j]){ #If the ith point_A falls within the jth start & end

      if(is.na(x$symbol[i])){ #If there is no symbol assigned to the ith row of x

        x$symbol[i] <- genes$symbol[j] #Assign the symbol from the jth row

      } else{ #If there is a symbol assigned to the ith row of x already, and it matches (now, another) jth row of genes
        x$symbol[i] <- paste(x$symbol[i], genes$symbol[j]) #Concatenate the new symbol from the jth row of genes to the ith row of x
      }

    }

  }
}

x

#   chr_1   point_A                         symbol
# 1     2 143300988 AC092578.1 Newadditionalsymbol
# 2     2 143611665                     AC013444.1
# 3     2 143635068                           KYNU
# 4     2 143704870             KYNU RP11-470B22.1
# 5     2 143763270                KYNU AC013444.2
# 6     2 143848932                       ARHGAP15

map gene positions to chromosome coordinates

5 Answers5