This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Merge 2 files based on rs number but keep all info only from the first file

I would like to merge two files by based on one column (rs number). Not all rs numbers in file 1 are in file 2, and not all in file 2 are in file 1. However, I wanna keep ALL lines from file 1 in the output.

File 1 (example):

rs    chr    bp    p
rs1    chr1    729679    0.9456
rs2    chr1    731718    0.7125
rs3    chr1    734349    0.7582
rs4    chr1    736289    0.7807
rs5    chr1    748878    0.3864

File 2 (example):

RSID CODING UTR PROMOTER DHS INTRON
rs2 0 0 0 0 1
rs3 0 0 0 0 1
rs1 0 0 1 1 0
rs4 0 0 0 1 0
rs6 0 1 0 1 0

In the case that an rs number is not found in file 2, I want to give this rs number the annotation: 0 0 0 0 0. so, hypothetically... based on my examples, this should be my output

Output (example):

rs    chr    bp    p    CODING    UTR    PROMOTER    DHS    INTRON
rs1    chr1    729679    0.9456    0    0    1    1    0
rs2    chr1    731718    0.7125    0    0    0    0    1
rs3    chr1    734349    0.7582    0    0    0    0    1
rs4    chr1    736289    0.7807    0    0    0    1    0
rs5    chr1    748878    0.3864    0    0    0    0    0

Merging files is fairly simple with a function like "join" but I have no idea how to keep the lines in file 1 without a column match with file 2, and in such a case, add the 0 0 0 0 0 annotation.

I really hope my question is clear, and any help is appreciated.

snp awk
• 5,582 views • link
updated by Ram Ram • written by joshf • joshf

5 answers

If you use Python, check out the pandas documentation for merging dataframes.

You might also consider writing a simple script:

• 0 views • link
written by Kamil Kamil

use join with option -a

join -t ' ' -1 1 -2 1 -a 1 <(sort -t ' ' -k1,1 input1.txt) <(sort -t ' ' -k1,1 input2.txt) | awk -F ' ' '{printf("%s",$0); if(NF==4) printf(" 0 0 0 0 0");printf("\n");}'

rs1 chr1 729679 0.9456 0 0 1 1 0
rs2 chr1 731718 0.7125 0 0 0 0 1
rs3 chr1 734349 0.7582 0 0 0 0 1
rs4 chr1 736289 0.7807 0 0 0 1 0
rs5 chr1 748878 0.3864 0 0 0 0 0
RSID chr bp p CODING UTR PROMOTER DHS INTRON
• 0 views • link
updated by Ram Ram • written by Pierre Lindenbaum Pierre Lindenbaum

I stumbled over a join command without the need of an extra awk call:

join -a 1 -t ' ' -e '0' -o '0,1.2,1.3,1.4,2.2,2.3,2.4,2.5' <(sort -t ' ' -k1,1 input1.txt) <(sort -t ' ' -k1,1 input2.txt)

It was said that the -e (Empty) option needs -o (output sequence). The 0 in -o is the joining field, followed by the file number and its column. Thus if file number 2 has a missing entry, the -e argument is put on each field defined in -o 2.X.

• 0 views • link
updated by Ram Ram • written by michael.ante michael.ante

In R with the help of dplyr package,

# 1.t and 2.t are your two files

x1 <- read.table("./1.t", header=TRUE)
x2 <- read.table("./2.t", header=TRUE)

x3 <- dplyr::left_join(x1, x2, by=c("rs"="RSID"))

# Change NA to 0
x3[whichis.na(x3), arr.ind=TRUE)]=0

Show x3:

   rs  chr     bp      p CODING UTR PROMOTER DHS INTRON
1 rs1 chr1 729679 0.9456      0   0        1   1      0
2 rs2 chr1 731718 0.7125      0   0        0   0      1
3 rs3 chr1 734349 0.7582      0   0        0   0      1
4 rs4 chr1 736289 0.7807      0   0        0   1      0
5 rs5 chr1 748878 0.3864      0   0        0   0      0
• 0 views • link
updated by Ram Ram • written by Zhilong Jia Zhilong Jia

Really fast with Python:

D = {}
f = open(FIRST_FILE, "r")
for l in f: D[ l.split("\t")[0].strip() ] = l.strip()
f.close()

f = open(SECOND_FILE, "r")
for l in f:
    rsid = l.split("\t")[0].strip()
try:
    print l.strip() + "\t" + D[rsid]
f.close()
• 0 views • link
updated by Ram Ram • written by matteof • matteof

and the awk alternative

awk '{
if (/^rs[0-9]/) { if (FNR==NR) f1[$1]=$0; else f2[$1]=$0 }
} END {
for (rs in f1) {
 if (f2[rs]) { print f1[rs],f2[rs] }
 else { print f1[rs],rs,0,0,0,0,0 }
}
}' file1.txt file2.txt \
| perl -pe 's/ +/\t/g' | cut -f-4,6- | sort -V
• 1 views • link
updated by Ram Ram • written by Jorge Amigo Jorge Amigo

Log in to answer this question.

More posts like this