Pfam database parser

pfam-parser.sh

#!/bin/bash

echo "Super Mega Hyper Pfam database parser started. Sit back and relax :)"
echo "Begin parsing " $1 " file.."

#remove old output
rm -rf parser_output
mkdir parser_output

#parse command line arguments
isParseOrganisms=false
for i in $@;
do
	if [[ $i =~ --organisms ]];
        then
                isParseOrganisms=true
		continue
        fi
	
	if $isParseOrganisms
	then
		organisms[$orgCount]=$i
		orgCount=$[$orgCount+1]
	fi
done

echo "Will extract groups with GF TP Family and split them among the following organisms:"
echo ${organisms[@]}

for i in ${organisms[@]}; 
do
	echo "<groups>" >> "parser_output/$i.txt"
done

cat $1 | while read inputLine; 
do
	#identify group ID by searching the string "#=GF ID"
	if [[ $inputLine =~ \#=GF\ ID.* ]];
	then
		groupId=$inputLine
	fi

	#determine whether this group is a family or not by searching the string "#=GF TP Family"
        if [[ $inputLine =~ \#=GF\ TP\ Family.* ]];
        then
                isFamily=true
        fi

	#determine the sequence number in this group by searching the string "#=GF SQ"
	if [[ $inputLine =~ \#=GF\ SQ.* ]];
	then
		seqNr=$inputLine
	fi

	#determine the Pfam-A accession number by searching the string "#=GF AC"
	if [[ $inputLine =~ \#=GF\ AC.* ]];
	then
		accessionNr=$inputLine
	fi
		
	#determine the Pfam family description by searching the string "#=GF DE"
	if [[ $inputLine =~ \#=GF\ DE.* ]];
	then
		familyDesc=$inputLine
	fi

	#determine the group comments by searching the string "#=GF CC"
	if [[ $inputLine =~ \#=GF\ CC.* ]];
	then
		comment=$inputLine
	fi

	#collect group sequences by searching the string "#=GS"
	if [[ $inputLine =~ \#=GS.* ]];
	then
		sequences[$seqCount]=$inputLine
		seqCount=$[$seqCount+1]

		#detect if the sequence belongs to one of the organisms we specified
		for i in "${organisms[@]}";
		do
			if [[ $inputLine =~ .*_$i/.* ]];
			then
				match=$(echo "${foundOrganisms[@]:0}" | grep -o $i)  
				if [[ -z $match ]]
				then 
					foundOrganisms=(${foundOrganisms[@]} $i)
				fi
				break
			fi
		done		
	fi

	#detect end of group token.
	if [[ $inputLine =~ // ]];
	then
		stockholmCount=$[$stockholmCount+1]

		#if the alignment belongs to a family we want to keep it
		if $isFamily
		then
			for i in "${foundOrganisms[@]}";
			do
				outfile="parser_output/$i.txt"
				echo "<group hash='$stockholmCount'>" >> $outfile
				echo "<id hash='$stockholmCount'>${groupId:2}</id>" >> $outfile
				echo "<accessionNr hash='$stockholmCount'>${accessionNr:2}</accessionNr>" >> $outfile
				echo "<description hash='$stockholmCount'>${familyDesc:2}</description>" >> $outfile
				echo "<comment hash='$stockholmCount'>${comment:2}</comment>" >> $outfile
				echo "<sequenceCount hash='$stockholmCount'>${seqNr:2}</sequenceCount>" >> $outfile
				echo "<sequences hash='$stockholmCount'>" >> $outfile
				for j in "${sequences[@]}";
				do
					#print only this organism's sequences
					if [[ $j =~ .*$i.* ]];
					then
						echo "<sequence>${j:2}</sequence>" >> $outfile
					fi
				done
				echo "</sequences>" >> $outfile
				echo "</group>" >> $outfile
			done
		fi

		#reset iteration variables
		unset alignment
		unset inputLineCount
		unset isFamily
		unset groupId
		unset accessionNr
		unset seqNr
		unset familyDesc
		unset comment
		unset organismCount
		unset seqcount
		unset fileCount
		unset foundOrganisms
		unset sequences

		echo "Processed group: $stockholmCount"
	fi
done

for i in ${organisms[@]}; 
do
        echo "</groups>" >> "parser_output/$i.txt"
done

To use the parser just download it to the same folder where you have your Pfam-A.full database and then type sh parser-sh Pfam-A.full --organisms [ORGANISM 1] [ORGANISM 2], etc..

It'll then create a folder named parser_output with a .txt file per organism. The output is one xml node per group. you can then use such output in a visual XML editor like xmlgrid.net. If your output files are just too big, I recommend splitting them into smaller ones