2

I've got two genetic datasets which have matching chromosome position IDs. I want to count how many times file 1's chromosome position IDs appear in file 2.

For example my data looks like:

File 1 (chromosome position is actually my 125th column, implied by the ...):

Gene  pval    ... Chromosome position ID
ACE   0.002   ... 01:3290834_CT_C_1
NOS   0.01    ... 03:3304593_GA_G_1
BRCA  0.004 . ... 06:6265733_GA_G_1
CYP3  0.34    ... 09:9433933_GA_G_1

File 2 (chromosome position is my first column):

Chromosome position ID  Gene  pval
01:1243933_GA_G_1       ACE   0.002
03:3304593_GA_G_1       NOS   0.01
06:6265733_GA_G_1       BRCA  0.004
09:9433933_GA_G_1       CYP3  0.34

I've found a lot of questions giving extraction of matching lines, and applied code based off those questions, but I just want to get the count of matching chromosome positions between 2 files.

Currently I'm using:

awk -F'|' 'NR==FNR{c[$125]++;next};c[$125]' file2.csv file1.txt > file3.txt

wc -l file1.txt
wc -l file3.txt

The line count for files 1 and 3 doesn't exactly match as I'm expecting (I expect all of file 1 to be in file 2) and to be sure what's going on I need to find a way to perform a count of matching rows in the chromosome position column. If I can find a way to code 'does all of file 1 chromosome positions match/appear in file 3?' that would be ideal, but even just a count works for now.

So the output would be a number counting how many times chromosome position column $125 in file 1 has matches also with file 2 chromosome position column $1.

I am using Linux.

6
  • 1
    Why are you using -F'|'? Your files don't even contain any |, at least none that you show. What defined a field in these files?
    – terdon
    Jan 31, 2020 at 11:15
  • Hi thank you for this, I will remove it. I'm new to linux, and just trying to piece this command together based on what I'm finding online.
    – DN1
    Jan 31, 2020 at 11:17
  • Ah, I see. The -F is setting the field delimiter, so when you use -F'|', it is expecting data separated into fields by a |, like foo|bar|baz. What separates the fields in your files? Is it tabs? Spaces? Something else?
    – terdon
    Jan 31, 2020 at 11:18
  • 1
    Some good example input there, but some example output would be useful too, in order to aid those trying to answer.
    – steve
    Jan 31, 2020 at 11:21
  • Also, could you please edit your question and make the title match the content? Your title is asking for something completely different (replacing a character), but your question doesn't seem to be involving any replacement at all.
    – terdon
    Jan 31, 2020 at 11:23

2 Answers 2

2

A solution with awk, tail, sort, join and wc.

join <(awk -F '\t' '{print $125}' file1 | tail -n +2 | sort) <(awk -F '\t' '{print $1}' file2 | tail -n +2 | sort ) | wc -l
3

Explanation.

This solution assumes that columns are tab separated. With awk are collected the 125th column of file1 and the first column of file2. tail -n +2 removes the first line of the collected results. sort is mandatory since join expects ordered files. The resulting intersecting set is supplied to wc which returns its number of lines.

5
  • Thank you, is the number 3 at the end meant to be a new file?
    – DN1
    Jan 31, 2020 at 11:59
  • In problem statement you said: the output would be a number counting how many times chromosome position column $125 in file 1 has matches also with file 2 chromosome position column $1. Jan 31, 2020 at 12:07
  • To see the resulting intersecting set remove | wc -l from the command. Jan 31, 2020 at 12:20
  • Thank you, this seems to work for me. Oddly it gives back 0 matches, which shouldn't be possible, but I'm wondering if this isn't a code problem anymore. One file is a text file and one is a csv, would that be an issue?
    – DN1
    Jan 31, 2020 at 13:00
  • If files have different separators that will be an issue. In the CVS file replace the awk separator by -F ','. Jan 31, 2020 at 13:03
1

You're close. Try

awk 'FNR == 1 {next}; FNR==NR {P[$125]; next} $1 in P {P[$1]++} END {for (p in P) print p, P[p]+0}' file[12]
03:3304593_GA_G_1 1
01:3290834_CT_C_1 0
09:9433933_GA_G_1 1
06:6265733_GA_G_1 1

Obviously, not all positions in file1 are found in file2.

2
  • Thank you, is the file[12] representative of both files 1 and 2? I currently have different file names in actuality but I can rename them if it works like this
    – DN1
    Jan 31, 2020 at 11:59
  • file[12] is expanded by the shell to file1 file2 and will be operated upon in that sequence by awk.
    – RudiC
    Jan 31, 2020 at 13:52

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .