CODEWITHSUNDEEP
rdataframesubset
André Conceição
André Conceição

Reputation: 13

Subseting by string

im working with a data.frame: on the rows I have probes names and in one colunm I have some iformation about the region where that probes are in the gene (1stExon, Body, etc), but I have a problem:

                  Gene          Gene_Region
cg14736058 PROM1;PROM1;PROM1  TSS200;5'UTR;1stExon
      .            .          1stExon;1stExon;1stExon;1stExon
      .            .          1stExon;1stExon;1stExon
      .            .          1stExon;1stExon;1stExon
      .            .          1stExon;1stExon;5'UTR;5'UTR;5'UTR;1stExon
      .            .          1stExon;1stExon
      .            .          1stExon;1stExon;Body
      .            .          Body;Body

I want the rows where only one region is present, but imanige imagine that "1sExon" is repetead but only that I want that row too. For example, I want the last row because "Body" is the only region repeated, so I consider it as one region only. I dont know if I am mking myself clear. PS. I dont know how many times a string is repeated

Upvotes: 0

Views: 58

Answers (1)

pyll
pyll

Reputation: 1764

This should do the trick. First collapse the string to only contain unique values. If there is only one unique value, the separator ; will disappear. So then you can just delete the rows that still contain a ;.

# Load Data
df <- structure(list(Gene_Region = c("TSS1500;5'UTR", "TSS1500;TSS1500;TSS1500;TSS1500", "Body", "1stExon;5'UTR", "1stExon;1stExon;1stExon", "Body", "Body;Body;Body;Body;Body" ), UCSC_RefGene_Name = c("USP44;USP44", "COL11A2;COL11A2;COL11A2;COL11A2", "SOX2OT", "CRYGD;CRYGD", "ENPP2;ENPP2;ENPP2", "PGLYRP2", "KCNQ2;KCNQ2;KCNQ2;KCNQ2;KCNQ2" )), .Names = c("Gene_Region", "UCSC_RefGene_Name"), row.names = c("cg13879483", "cg08481075", "cg13294849", "cg22399133", "cg02534163", "cg16206460", "cg13782274"), class = "data.frame")

# Collapse
df$unique_regions <- sapply(strsplit(df$Gene_Region ,";"), function(x) 
  paste(unique(x), collapse=";"))

# Search for rows with no semicolon in unique_regions column
df2 <- df[- grep(";", df$unique_regions),]

Upvotes: 2

Related Questions