3

I have been running roary pangenome pipeline, where I need to write a script in for loop, For example I have a gff files as follows,

a.gff
b.gff
5.gff
101.gff
clustered_proteins

I need to run a command for retrieving unique genes from roary pipeline as follows,

query_pan_genome -a difference --input_set_one a.gff --input_set_two b.gff,5.gff,101.gff -g clustered_proteins
 
query_pan_genome -a difference --input_set_one b.gff --input_set_two a.gff,5.gff,101.gff -g clustered_proteins

query_pan_genome -a difference --input_set_one 5.gff --input_set_two a.gff,b.gff,101.gff -g clustered_proteins

query_pan_genome -a difference --input_set_one 101.gff --input_set_two a.gff,b.gff,5.gff -g clustered_proteins

For doing the same I wrote a script as follows,

file1=*.gff
file2=*.gff
file3="-f "$file1-$file2"
for file in *.gff
do
query_pan_genome -a difference --input_set_one "$file1" --input_set_two "$file3" -g clustered_proteins 
done

But above script is not serving my purpose, I knew simple for script only, this is something difficult for me to write. Kindly help me to make the script perfect.

Thanks in advance.

Improve this question

3 Answers 3

Reset to default
6

Using /bin/sh:

#!/bin/sh

set -- *.gff
for name do
    shift
    (   IFS=,
        query_pan_genome -a difference \
            --input_set_one "$name" \
            --input_set_two "$*" \
            -g clustered_proteins
    )
    set -- "$@" "$name"
done

This first sets the positional parameters to the list of filenames matching the pattern *.gff. It then loops over this list, setting the variable name to the value of the current filename.

In the loop, for each filename, the first positional parameter is shifted off of the list of positional parameters. This shifted off element corresponds to the current filename also found in $name. After the shift, the list of positional parameters contains the other filenames, but not $name.

In the (...) subshell in the loop, we locally set IFS to a comma. This means that "$*" would be expanded to a string consisting of the current list of positional parameters with a comma as delimiter.

We then call the query_pan_genome utility with the $name filename and the others as a comma-delimited list.

After the subshell, just before the end of the body of the loop, the current name is put back into the list of positional parameters, but at the end.

There is no issue with modifying the list of positional parameters inside the loop, even if we're iterating over this list. The list that we iterate over is an unchanging copy of the list that we modify in the loop with shift and set (for loops always iterate over a static list of elements).

Given the list of GFF files in the question, the code above would end up executing the following four commands:

query_pan_genome -a difference --input_set_one 101.gff --input_set_two 5.gff,a.gff,b.gff -g clustered_proteins
query_pan_genome -a difference --input_set_one 5.gff --input_set_two a.gff,b.gff,101.gff -g clustered_proteins
query_pan_genome -a difference --input_set_one a.gff --input_set_two b.gff,101.gff,5.gff -g clustered_proteins
query_pan_genome -a difference --input_set_one b.gff --input_set_two 101.gff,5.gff,a.gff -g clustered_proteins
Improve this answer
5

It's easier with zsh:

#! /bin/zsh -
files=(*.gff(N))

# don't run that command if there are fewer than 2 files
(( $#files < 2 )) ||
  for f ($files)
    query_pan_genome -a difference \
                     --input_set_one $f \
                     --input_set_two ${(j[,])files:#$f} \
                     -g clustered_proteins

Where

  • ${array:#pattern} expands to the elements of the array that don't match the pattern, so here with ${files#$f} the elements of $files except $f.
  • ${(j[,])array} joins the elements of the array with ,.

Instead of *.gff(N), you may want to use *.gff(Nn) where the n glob qualifier turns on the numericglobsort option for the expansion of that one glob, so that file10.gff comes after file2.gff for instance.

Improve this answer
0
5

You can use an array, looping over its indices and temporarily unsetting each index:

#! /bin/bash

input_files=(*.gff)

IFS=,
for index in "${!input_files[@]}"
do
    input_file=${input_files[$index]}
    unset input_files[$index]
    echo "$input_file" "${input_files[*]}"
    input_files[$index]=$input_file
done

Example output:

101.gff 5.gff,a.gff,b.gff
5.gff 101.gff,a.gff,b.gff
a.gff 101.gff,5.gff,b.gff
b.gff 101.gff,5.gff,a.gff
Improve this answer
2
  • Thank you @muru, Could you please elaborate how to use your code in my roary pipeline command. I do not know how to incorporate your code in my roary pipeline command. Moreover, when I execute your code as its, I am getting error as follows, ga@ga214:~/Documents/data/xap98/gff_gbk_dna_protein/gff$ sh uniq.sh uniq.sh: 3: uniq.sh: Syntax error: "(" unexpected Commented Feb 24, 2021 at 5:25
  • @Kumar instead of echo "$input_file" "${input_files[*]}" you use your query_pan_genome command with the flags and appropriate variables. This is a bash script, run it with bash, not sh. Commented Feb 24, 2021 at 5:32

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.