CODEWITHSUNDEEP
pythonbashparsingawksed
Geroge
Geroge

Reputation: 561

Parsing data in awk

I have a genetics data like this:

MUT1    G_->_A_(het)    44%_(96)___[45%_(49)_/_43%_(47)]    rs1799967_(Gene_file;_1000Genomes;_ClinVarVCF;_dbSNP,MutDB) c.4956G>A   1
MUT1    A_->_G_(homo)   99%_(297)___[99%_(151)_/_99%_(146)] rs206075_(Gene_file;_1000Genomes;_ClinVarVCF;_dbSNP)    c.4563A>G   1
MUT1    G_->_C_(homo)   100%_(259)___[100%_(132)_/_100%_(127)]  COSM4147689_(COSMIC),_COSM4147690_(COSMIC),_rs206076_(Gene_file;_1000Genomes;_ClinVar;_ClinVarVCF;_dbSNP)   c.6513G>C   2
MUT1    A_->_C_(het)    41%_(103)___[42%_(53)_/_40%_(50)]   COSM3753646_(COSMIC),_COSM147663_(COSMIC),_rs144848_(Gene_file;_1000Genomes;_ClinVarVCF;_dbSNP,MutDB)   c.1114A>C   5

What I need to parse this data and extract only some kind of fields.

Require output is:

MUT1    het 44% rs1799967 c.4956G>A 1
MUT1    homo 99% rs206075c.4563A>G  1
MUT1    homo 100% rs206076 c.6513G>C    2
MUT1    het 41% rs144848 c.1114A>C  5

So output shoould be - all first column, from second column only het or hom, third column is only %, fifth column should be extracted only rs_number - this have always different position and last column.

note: I know, that info about homo/het is always in last filed of second column. And % is always on first field in third column.

My solution is:

awk -v OFS="\t" '{print $1,$5,$6,$9,$10,$11}' zkouska.csv | awk -v OFS="\t" 'NR>1{split($2,arr2,"_"); split($3,arr3,"_"); print $1,arr2[4],arr3[1],$4,$5,$6}'

But output is :

BRCA1   (het)   44% rs1799967_(Gene_file;_1000Genomes;_ClinVarVCF;_dbSNP,MutDB) c.4956G>A   1
BRCA1   (homo)  99% rs206075_(Gene_file;_1000Genomes;_ClinVarVCF;_dbSNP)    c.4563A>G   1
BRCA1   (homo)  100%    COSM4147689_(COSMIC),_COSM4147690_(COSMIC),_rs206076_(Gene_file;_1000Genomes;_ClinVar;_ClinVarVCF;_dbSNP)   c.6513G>C   2
BRCA1   (het)   41% COSM3753646_(COSMIC),_COSM147663_(COSMIC),_rs144848_(Gene_file;_1000Genomes;_ClinVarVCF;_dbSNP,MutDB)   c.1114A>C   5
BRCA1   (homo)  100%    COSM148277_(COSMIC),_COSM3755561_(COSMIC),_rs16942_(Gene_file;_1000Genomes;_ClinVarVCF;_dbSNP)  c.3548A>G   5

Still have problem to extract rs from fifth column. Erase quotes in second field. Inpout and output should be TAB separate. Solution could no be only in awk.

Upvotes: 3

Views: 390

Answers (3)

Inian
Inian

Reputation: 85690

Am posting a pure bash logic for your requirement.

#!/bin/bash

while read col1 col2 col3 col4 col5 col6
do
    subcol2="${col2#*(}";subcol2=${subcol2%)*}                             # Extracting string within the braces '()' using parameter-expansion              
    [[ $col4 =~ .*rs([[:digit:]]+).* ]] && subcol4="${BASH_REMATCH[1]}"    # RegEx in bash to extract number following the 'rs' string
    printf "%s %s %s %s %s %s\n" "$col1" "$subcol2" "${col3%%_*}" "rs$subcol4" "$col5" "$col6"
done <file

On running the script, produces the result as

$ bash script.sh
MUT1 het 44% rs1799967 c.4956G>A 1
MUT1 homo 99% rs206075 c.4563A>G 1
MUT1 homo 100% rs206076 c.6513G>C 2
MUT1 het 41% rs144848 c.1114A>C 5

NOTE: The solution may perform slower on bigger files. I have tested this only on your sample file.

Upvotes: 3

Thor
Thor

Reputation: 47119

Using a combination of gsub and match might be the way to go, here is a portable example:

parse.awk

{
  gsub(/^[^(]+\(|\)/, "", $2)
  gsub(/_.*/, "", $3)
  match($4, /rs[0-9]+/)
  print $1, $2, $3, substr($4, RSTART, RLENGTH), $5, $6
}

Run it like this:

awk -f parse.awk OFS='\t' < infile

Output:

MUT1    het     44%     rs1799967   c.4956G>A   1
MUT1    homo    99%     rs206075    c.4563A>G   1
MUT1    homo    100%    rs206076    c.6513G>C   2
MUT1    het     41%     rs144848    c.1114A>C   5

Upvotes: 3

Sundeep
Sundeep

Reputation: 23677

$ perl -lne 'print join "\t", /^(\S+)/,/^[^(]+\(\K([^)]+)/,/^[^)]+\)\s+\K(\d+%)/,/(rs\d+)/,/(\S+\s+\S+)\s*$/' file
MUT1    het 44% rs1799967   c.4956G>A   1
MUT1    homo    99% rs206075    c.4563A>G   1
MUT1    homo    100%    rs206076    c.6513G>C   2
MUT1    het 41% rs144848    c.1114A>C   5
  • /^(\S+)/ extract non-whitespace characters from start of line
  • /^[^(]+\(\K([^)]+)/ extract characters between first ()
  • /^[^)]+\)\s+\K(\d+%)/ extract first match of digits followed by % after first ) in the line
  • /(rs\d+)/ extract rs followed by digits
  • /(\S+\s+\S+)\s*$/ extract last two columns


Another way is to process each field separately, similar to bash and awk solutions

$ perl -lane '
$F[1] =~ s/.*\(|\)//g;
$F[2] =~ s/_.*//;
($F[3]) = $F[3] =~ m/(rs\d+)/;
print join "\t", @F;
' file
MUT1    het 44% rs1799967   c.4956G>A   1
MUT1    homo    99% rs206075    c.4563A>G   1
MUT1    homo    100%    rs206076    c.6513G>C   2
MUT1    het 41% rs144848    c.1114A>C   5

Upvotes: 4

Related Questions