view precursors.pl @ 20:d88fc3c0084e draft

Uploaded
author big-tiandm
date Fri, 25 Jul 2014 05:32:24 -0400
parents 1131b4008650
children 4c0b1a94b882
line wrap: on
line source

#!/usr/bin/perl -w
#Filename:
#Author: Tian Dongmei
#Email: tiandm@big.ac.cn
#Date: 2013/7/19
#Modified:
#Description: 
my $version=1.00;

use strict;
use Getopt::Long;
use RNA;

my %opts;
GetOptions(\%opts,"map=s","g=s","d:i","f:i","o=s","e:f","s=s","h");
if (!(defined $opts{map} and defined $opts{g} and defined $opts{o} and defined $opts{s} ) || defined $opts{h}) { #necessary arguments
&usage;
}

my $filein=$opts{'map'};
my $faout=$opts{'o'};
my $strout=$opts{'s'};
my $genome= $opts{'g'};

my $maxd=defined $opts{'d'} ? $opts{'d'} : 200;
my $flank=defined $opts{'f'}? $opts{'f'} : 10;

my $MAX_ENERGY=-18;
if (defined $opts{'e'}) {$MAX_ENERGY=$opts{'e'};}
my $MAX_UNPAIR=5;
my $MIN_PAIR=15;
my $MAX_SIZEDIFF=4;
my $MAX_BULGE=2;
my $ASYMMETRY=5;
my $MIN_UNPAIR=0;
my $MIN_SPACE=5;
my $MAX_SPACE=$maxd;
my $FLANK=$flank;

######### load in genome sequences start ########
my %genome;
my %lng;
my $name;
open IN,"<$genome";
while (my $aline=<IN>) {
	chomp $aline;
	next if($aline=~/^\#/);
	if ($aline=~/^>(\S+)/) {
		$name=$1;
		next;
	}
	$genome{$name} .=$aline;
}
close IN;
foreach my $key (keys %genome) {
	$lng{$key}=length($genome{$key});
}
####### load in genome sequences end ##########

my %breaks; ### reads number bigger than 3
open IN,"<$filein"; #input file  
while (my $aline=<IN>) {
	chomp $aline;
	my @tmp=split/\t/,$aline;
	$tmp[0]=~/_x(\d+)$/;
	my $no=$1;
	next if($no<3);
	#my $trand=&find_strand($tmp[9]);
	#my @pos=split/\.\./,$tmp[5];
	my $end=$tmp[3]+length($tmp[4])-1;
	if($tmp[1] eq "-"){$tmp[4]=revcom($tmp[4]);}
	push @{$breaks{$tmp[2]}{$tmp[1]}},[$tmp[3],$end,$no,$tmp[4]]; ### 0 base
}
close IN;

my %cites; ### peaks
foreach my $chr (keys %breaks) {
	foreach my $strand (keys %{$breaks{$chr}}) {
		my @array=@{$breaks{$chr}{$strand}};
		@array=sort{$a->[0]<=>$b->[0]} @array;
		for (my $i=0;$i<@array;$i++) {
			my $start=$array[$i][0];my $end=$array[$i][1];
			my @subarray=();
			push @subarray,$array[$i];

			for (my $j=$i+1;$j<@array;$j++) {
				if ($start<$array[$j][1] && $end>$array[$j][0]) {
					push @subarray,$array[$j];
					($start,$end)=&newpos($start,$end,$array[$j][0],$array[$j][1]);
				}
				else{
					$i=$j;
					&find_cites(\@subarray,$chr,$strand);
					last;
				}
			}
		}
	}
}

open FA,">$faout"; #output file 
open STR,">$strout";
foreach my $chr (keys %cites) {
	foreach my $strand (keys %{$cites{$chr}}) {

		my @array2=@{$cites{$chr}{$strand}};
		@array2=sort{$a->[0]<=>$b->[0]} @array2;
		&excise(\@array2,$chr,$strand);
	}
}
close FA;
close STR;
sub oneCiteDn{
	my ($array,$a,$chr,$strand)=@_;

	my $ss=$$array[$a][0]-$flank;
	$ss=0 if($ss<0);
	my $ee=$$array[$a][1]+$maxd+$flank;
	$ee=$lng{$chr} if($ee>$lng{$chr});
	
	my $seq=substr($genome{$chr},$ss,$ee-$ss+1);
	if($strand eq "-"){$seq=revcom($seq);}

	my $val=&ffw1($seq,$$array[$a][3],$chr,$strand,$ss,$ee);
	return $val;
}
sub oneCiteUp{
	my ($array,$a,$chr,$strand)=@_;

	my $ss=$$array[$a][0]-$maxd-$flank;
	$ss=0 if($ss<0);
	my $ee=$$array[$a][1]+$flank;
	$ee=$lng{$chr} if($ee>$lng{$chr});
	
	my $seq=substr($genome{$chr},$ss,$ee-$ss+1);
	if($strand eq "-"){$seq=revcom($seq);}

	my $val=&ffw1($seq,$$array[$a][3],$chr,$strand,$ss,$ee);
	return $val;

}

sub twoCites{
	my ($array,$a,$b,$chr,$strand)=@_;

	my $ss=$$array[$a][0]-$flank;
	$ss=0 if($ss<0);
	my $ee=$$array[$b][1]+$flank;
	$ee=$lng{$chr} if($ee>$lng{$chr});
	
	my $seq=substr($genome{$chr},$ss,$ee-$ss+1);
	if($strand eq "-"){$seq=revcom($seq);}

#	my( $str,$mfe)=RNA::fold($seq);
#	return 0 if($mfe>$MAX_ENERGY); ### minimum mfe
	my $val=&ffw2($seq,$$array[$a][3],$$array[$b][3],$chr,$strand,$ss,$ee);
	
	return $val;

}
sub excise{
	my ($cluster,$chr,$strand)=@_;

	my $last_pos=0;
	for (my $i=0;$i<@{$cluster};$i++) {
		next if($$cluster[$i][0]<$last_pos);
		my $ok=0;
		for (my $j=$i+1;$j<@{$cluster} ;$j++) {
			if($$cluster[$j][0]-$$cluster[$i][1]>$maxd){
				$i=$j;
				last;
			}else{
				$ok=&twoCites($cluster,$i,$j,$chr,$strand);
				if($ok){ $last_pos=$$cluster[$j][1]+$flank; $i=$j; last;}
			}
		}
		next if($ok);

		$ok=&oneCiteDn($cluster,$i,$chr,$strand);
		if($ok){$last_pos=$$cluster[$i][1]+$maxd+$flank; next;}
		$ok=&oneCiteUp($cluster,$i,$chr,$strand);
		if($ok){$last_pos=$$cluster[$i][1]+$flank;next;}
	}


}

sub ffw2{
	my ($seq,$tag1,$tag2,$chr,$strand,$ss,$ee)=@_;
	
	my $N_count=$seq=~tr/N//;  ## precursor sequence has not more than 5 Ns
	if ($N_count > 5) {
		return 0;
	}

	my $seq_length=length $seq;
	# position tag1 and tag2
	my $tag1_beg=index($seq,$tag1,0)+1;
	if ($tag1_beg < 1) {
		warn "[ffw2] coordinate error.\n";
#		$fold->{reason}="coordinate error";
		return 0;
	}
	my $tag2_beg=index($seq,$tag2,0)+1;
	if ($tag2_beg < 1) {
		warn "[ffw2] coordinate error.\n";
#		$fold->{reason}="coordinate error";
		return 0;
	}
	if ($tag2_beg < $tag1_beg) {
		# swap tag1 and tag2
		($tag1,$tag2)=($tag2,$tag1);
		($tag1_beg,$tag2_beg)=($tag2_beg,$tag1_beg);
	}
	my $tag1_end=$tag1_beg+length($tag1)-1;
	my $tag2_end=$tag2_beg+length($tag2)-1;
	# re-clipping
	my $beg=$tag1_beg-$FLANK; $beg=1 if $beg < 1;
	my $end=$tag2_end+$FLANK; $end=$seq_length if $end > $seq_length;
	$seq=substr($seq,$beg-1,$end-$beg+1);
	$seq_length=length $seq;
	# re-reposition
	$tag1_beg=index($seq,$tag1,0)+1;
	if ($tag1_beg < 1) {
		warn "[ffw2] coordinate error.\n";
#		$fold->{reason}="coordinate error";
		return 0;
	}
	
	$tag2_beg=index($seq,$tag2,0)+1;
	if ($tag2_beg < 1) {
		warn "[ffw2] coordinate error.\n";
#		$fold->{reason}="coordinate error";
		return 0;
	}
	$tag1_end=$tag1_beg+length($tag1)-1;
	$tag2_end=$tag2_beg+length($tag2)-1;

	# fold
	my ($struct,$mfe)=RNA::fold($seq);
	$mfe=sprintf "%.2f", $mfe;
	if ($mfe > $MAX_ENERGY) {return 0;}

	# tag1
	my $tag1_length=$tag1_end-$tag1_beg+1;
	my $tag1_struct=substr($struct,$tag1_beg-1,$tag1_length);
	my $tag1_arm=which_arm($tag1_struct);
	my $tag1_unpair=$tag1_struct=~tr/.//;
	my $tag1_pair=$tag1_length-$tag1_unpair;
	my $tag1_max_bulge=biggest_bulge($tag1_struct);
	if ($tag1_arm ne "5p") { return 0;} # tag not in stem
#	if ($tag1_unpair > $MAX_UNPAIR) {$fold->{reason}="unpair=$tag1_unpair ($MAX_UNPAIR)"; return $pass}
	if ($tag1_pair < $MIN_PAIR) {return 0;}
	if ($tag1_max_bulge > $MAX_BULGE) {return 0;}

	# tag2
	my $tag2_length=$tag2_end-$tag2_beg+1;
	my $tag2_struct=substr($struct,$tag2_beg-1,$tag2_length);
	my $tag2_arm=which_arm($tag2_struct);
	my $tag2_unpair=$tag2_struct=~tr/.//;
	my $tag2_pair=$tag2_length-$tag2_unpair;
	my $tag2_max_bulge=biggest_bulge($tag2_struct);
	if ($tag2_arm ne "3p") {return 0;} # star not in stem
#	if ($tag2_unpair > $MAX_UNPAIR) {$fold->{reason}="unpair=$tag2_unpair ($MAX_UNPAIR)"; return $pass}
	if ($tag2_pair < $MIN_PAIR) {return 0;}
	if ($tag2_max_bulge > $MAX_BULGE) {return 0;}

	# space size between miR and miR*
	my $space=$tag2_beg-$tag1_end-1;
	if ($space < $MIN_SPACE) {return 0;}
	if ($space > $MAX_SPACE) {return 0;}
	
	# size diff of miR and miR*
	my $size_diff=abs($tag1_length-$tag2_length);
	if ($size_diff > $MAX_SIZEDIFF) {return 0;}

	# build base pairing table
	my %pairtable;
	&parse_struct($struct,\%pairtable); # coords count from 1

	my $asy1=get_asy(\%pairtable,$tag1_beg,$tag1_end);
	my $asy2=get_asy(\%pairtable,$tag2_beg,$tag2_end);
	my $asy=($asy1 < $asy2) ? $asy1 : $asy2;
	if ($asy > $ASYMMETRY) {return 0}

	# duplex fold, determine whether two matures like a miR/miR* ike duplex
	my ($like_mir_duplex1,$duplex_pair,$overhang1,$overhang2)=likeMirDuplex1($tag1,$tag2);
	# parse hairpin, determine whether two matures form miR/miR* duplex in hairpin context
	my ($like_mir_duplex2,$duplex_pair2,$overhang_b,$overhang_t)=likeMirDuplex2(\%pairtable,$tag1_beg,$tag1_end,$tag2_beg,$tag2_end);
	if ($like_mir_duplex1==0 && $like_mir_duplex2==0) {
		return 0;
	}

	print FA ">$chr:$strand:$ss..$ee\n$seq\n";
	print STR ">$chr:$strand:$ss..$ee\n$seq\n$struct\t($mfe)\n";
	
	return 1;	
}

sub ffw1{
	my ($seq,$tag,$chr,$strand,$ss,$ee)=@_;
	my $pass=0;

	my $N_count=$seq=~tr/N//;
	if ($N_count > 5) {
		return 0;
	}

	my $seq_length=length $seq;
	my $tag_length=length $tag;

	# position
	my $tag_beg=index($seq,$tag,0)+1;
	if ($tag_beg < 1) {
		 warn "[ffw1] coordinate error.\n";
		 return $pass;
	}
	my $tag_end=$tag_beg+length($tag)-1;
	
	
	# define candidate precursor by hybrid short arm to long arm, not solid enough
	my($beg,$end)=define_precursor($seq,$tag);
	if (not defined $beg) {
		return $pass;
	}
	if (not defined $end) {
		return $pass;
	}
	$seq=substr($seq,$beg-1,$end-$beg+1);
	$seq_length=length $seq;
	
	
	# fold
	my ($struct,$mfe)=RNA::fold($seq);
	$mfe=sprintf "%.2f",$mfe;
	if ($mfe > $MAX_ENERGY) {
		$pass=0;
		return $pass;
	}

	# reposition
	$tag_beg=index($seq,$tag,0)+1;
	if ($tag_beg < 1) {
		 warn "[ffw1] coordinate error.\n";
		 return 0;
	}
	$tag_end=$tag_beg+length($tag)-1;

	my $tag_struct=substr($struct,$tag_beg-1,$tag_length);
	my $tag_arm=which_arm($tag_struct);
	my $tag_unpair=$tag_struct=~tr/.//;
	my $tag_pair=$tag_length-$tag_unpair;
	my $tag_max_bulge=biggest_bulge($tag_struct);
	if ($tag_arm eq "-") { return $pass;}
#	if ($tag_unpair > $MAX_UNPAIR) {$fold->{reason}="unpair=$tag_unpair ($MAX_UNPAIR)"; return $pass}
	if ($tag_pair < $MIN_PAIR) { return $pass;}
	if ($tag_max_bulge > $MAX_BULGE) {return $pass;}

	# build base pairing table
	my %pairtable;
	&parse_struct($struct,\%pairtable); # coords count from 1
	
	# get star
	my ($star_beg,$star_end)=get_star(\%pairtable,$tag_beg,$tag_end);
	my $star=substr($seq,$star_beg-1,$star_end-$star_beg+1);
	my $star_length=$star_end-$star_beg+1;
	my $star_struct=substr($struct,$star_beg-1,$star_end-$star_beg+1);
	my $star_arm=which_arm($star_struct);
	my $star_unpair=$star_struct=~tr/.//;
	my $star_pair=$star_length-$star_unpair;
	my $star_max_bulge=biggest_bulge($star_struct);
	if ($star_arm eq "-") { return $pass;}
#	if ($star_unpair > $MAX_UNPAIR) {$fold->{reason}="unpair=$star_unpair ($MAX_UNPAIR)"; return $pass}
	if ($star_pair < $MIN_PAIR) {return $pass;}
	if ($star_max_bulge > $MAX_BULGE) {return $pass;}

	if ($tag_arm eq $star_arm) {return $pass;}
	
	# space size between miR and miR*
	my $space;
	if ($tag_beg < $star_beg) {
		$space=$star_beg-$tag_end-1;
	}
	else {
		$space=$tag_beg-$star_end-1;
	}
	if ($space < $MIN_SPACE) { return $pass;}
	if ($space > $MAX_SPACE) { return $pass;}

	# size diff
	my $size_diff=abs($tag_length-$star_length);
	if ($size_diff > $MAX_SIZEDIFF) { return $pass;}
	
	# asymmetry
	my $asy=get_asy(\%pairtable,$tag_beg,$tag_end);
	if ($asy > $ASYMMETRY) {return $pass;}
	
	$pass=1;
	print FA ">$chr:$strand:$ss..$ee\n$seq\n";
	print STR ">$chr:$strand:$ss..$ee\n$seq\n$struct\t($mfe)\n";
	return $pass;

}
sub get_star {
	my($table,$beg,$end)=@_;
	
	my ($s1,$e1,$s2,$e2); # s1 pair to s2, e1 pair to e2
	foreach my $i ($beg..$end) {
		if (defined $table->{$i}) {
			my $j=$table->{$i};
			$s1=$i;
			$s2=$j;
			last;
		}
	}
	foreach my $i (reverse ($beg..$end)) {
		if (defined $table->{$i}) {
			my $j=$table->{$i};
			$e1=$i;
			$e2=$j;
			last;
		}
	}
#	print "$s1,$e1 $s2,$e2\n";
	
	# correct terminus
	my $off1=$s1-$beg;
	my $off2=$end-$e1;
	$s2+=$off1;
	$s2+=2; # 081009
	$e2-=$off2; $e2=1 if $e2 < 1;
	$e2+=2; $e2=1 if $e2 < 1; # 081009
	($s2,$e2)=($e2,$s2) if ($s2 > $e2);
	return ($s2,$e2);
}

sub define_precursor {
	my $seq=shift;
	my $tag=shift;

	my $seq_length=length $seq;
	my $tag_length=length $tag;
	my $tag_beg=index($seq,$tag,0)+1;
	my $tag_end=$tag_beg+$tag_length-1;

	# split the candidate region into short arm and long arm
	my $tag_arm;
        my ($larm,$larm_beg,$larm_end);
	my ($sarm,$sarm_beg,$sarm_end);
        if ($tag_beg-1 < $seq_length-$tag_end) { # on 5' arm
                $sarm=substr($seq,0,$tag_end);
		$larm=substr($seq,$tag_end);
		$sarm_beg=1;
		$sarm_end=$tag_end;
		$larm_beg=$tag_end+1;
		$larm_end=$seq_length;
		$tag_arm="5p";
        }
        else {
            $larm=substr($seq,0,$tag_beg-1); # on 3' arm
			$sarm=substr($seq,$tag_beg-1);
			$larm_beg=1;
			$larm_end=$tag_beg-1;
			$sarm_beg=$tag_beg;
			$sarm_end=$seq_length;
			$tag_arm="3p";
        }
	
#	print "$sarm_beg,$sarm_end $sarm\n";
#	print "$larm_beg,$larm_end $larm\n";

	# clipping short arm
	if ($tag_arm eq "5p") {
		$sarm_beg=$tag_beg-$flank; $sarm_beg=1 if $sarm_beg < 1;
		$sarm=substr($seq,$sarm_beg-1,$sarm_end-$sarm_beg+1);
	}
	else {
		$sarm_end=$tag_end+$flank; $sarm_end=$seq_length if $sarm_end > $seq_length;
		$sarm=substr($seq,$sarm_beg-1,$sarm_end-$sarm_beg+1);
	}
#	print "$sarm_beg,$sarm_end $sarm\n";
#	print "$larm_beg,$larm_end $larm\n";

	# define the precursor by hybriding short arm to long arm
	my $duplex=RNA::duplexfold($sarm,$larm);
	my $struct=$duplex->{structure};
	my $energy=sprintf "%.2f", $duplex->{energy};
	my ($str1,$str2)=split(/&/,$struct);
	my $pair=$str1=~tr/(//;
#	print "pair=$pair\n";
	my $beg1=$duplex->{i}+1-length($str1);
	my $end1=$duplex->{i};
	my $beg2=$duplex->{j};
	my $end2=$duplex->{j}+length($str2)-1;
#	print "$beg1:$end1 $beg2:$end2\n";
	# transform coordinates
	$beg1=$beg1+$sarm_beg-1;
	$end1=$end1+$sarm_beg-1;
	$beg2=$beg2+$larm_beg-1;
	$end2=$end2+$larm_beg-1;
#	print "$beg1:$end1 $beg2:$end2\n";
	
	my $off5p=$beg1-$sarm_beg;
	my $off3p=$sarm_end-$end1;
	$beg2-=$off3p; $beg2=1 if $beg2 < 1;
	$end2+=$off5p; $end2=$seq_length if $end2 > $seq_length;

#	print "$beg1:$end1 $beg2:$end2\n";
	
	my $beg=$sarm_beg < $beg2 ? $sarm_beg : $beg2;
	my $end=$sarm_end > $end2 ? $sarm_end : $end2;
	
	return if $pair < $MIN_PAIR;
#	print "$beg,$end\n";
	return ($beg,$end);
}


# duplex fold, judge whether two short seqs like a miRNA/miRNA* duplex
sub likeMirDuplex1 {
	my $seq1=shift;
	my $seq2=shift;
	my $like_mir_duplex=1;
	
	my $length1=length $seq1;
	my $length2=length $seq2;
	my $duplex=RNA::duplexfold($seq1, $seq2);
	my $duplex_struct=$duplex->{structure};
	my $duplex_energy=sprintf "%.2f", $duplex->{energy};
	my ($str1,$str2)=split(/&/,$duplex_struct);
	my $beg1=$duplex->{i}+1-length($str1);
	my $end1=$duplex->{i};
	my $beg2=$duplex->{j};
	my $end2=$duplex->{j}+length($str2)-1;
	
	# revise beg1, end1, beg2, end2
	$str1=~/^(\.*)/;
	$beg1+=length($1);
	$str1=~/(\.*)$/;
	$end1-=length($1);
	$str2=~/^(\.*)/;
	$beg2+=length($1);
	$str2=~/(\.*)$/;
	$end2-=length($1);

	my $pair_num=$str1=~tr/(//;
	my $overhang1=($length2-$end2)-($beg1-1); # 3' overhang at hairpin bottom
	my $overhang2=($length1-$end1)-($beg2-1); # 3' overhang at hairpin neck
#	print $pair_num,"\n";
#	print $overhang1,"\n";
#	print $overhang2,"\n";
	if ($pair_num < 13) {
		$like_mir_duplex=0;
	}
	if ($overhang1 < 0 || $overhang2 < 0 ) {
		$like_mir_duplex=0;
	}
	if ($overhang1 > 4 || $overhang2 > 4) {
		$like_mir_duplex=0;
	}
	return ($like_mir_duplex,$pair_num,$overhang1,$overhang2);
}

# judge whether two matures form miR/miR* duplex, in hairpin context
sub likeMirDuplex2 {
	my ($table,$beg1,$end1,$beg2,$end2)=@_;
	my $like_mir_duplex=1;

#	   s1         e1
#   5 ----------------------------3	
#      | | |||| |||               |
#3 -------------------------------5
#      e2         s2

	my $pair_num=0;
	my $overhang1=0;
	my $overhang2=0;
	my ($s1,$e1,$s2,$e2);
	foreach my $i ($beg1..$end1) {
		if (defined $table->{$i}) {
			my $j=$table->{$i};
			if ($j <= $end2 && $j >= $beg2) {
				$s1=$i;
				$e2=$j;
				last;
			}
		}
	}
	foreach my $i (reverse ($beg1..$end1)) {
		if (defined $table->{$i}) {
			my $j=$table->{$i};
			if ($j <= $end2 && $j >= $beg2) {
				$e1=$i;
				$s2=$j;
				last;
			}
		}
	}

#	print "$beg1,$end1 $s1,$e1\n";
#	print "$beg2,$end2 $s2,$e2\n";

	foreach my $i ($beg1..$end1) {
		if (defined $table->{$i}) {
			my $j=$table->{$i};
			if ($j <= $end2 && $j >= $beg2) {
				++$pair_num;
			}
		}
	}
	if (defined $s1 && defined $e2) {
		$overhang1=($end2-$e2)-($s1-$beg1);
	}
	if (defined $e1 && defined $s2) {
		$overhang2=($end1-$e1)-($s2-$beg2);
	}
	
	if ($pair_num < 13) {
		$like_mir_duplex=0;
	}
	if ($overhang1 < 0 && $overhang2 < 0) {
		$like_mir_duplex=0;
	}
	return ($like_mir_duplex,$pair_num,$overhang1,$overhang2);
}
sub parse_struct {
	my $struct=shift;
	my $table=shift;

	my @t=split('',$struct);
	my @lbs; # left brackets
	foreach my $k (0..$#t) {
		if ($t[$k] eq "(") {
			push @lbs, $k+1;
		}
		elsif ($t[$k] eq ")") {
			my $lb=pop @lbs;
			my $rb=$k+1;
			$table->{$lb}=$rb;
			$table->{$rb}=$lb;
		}
	}
	if (@lbs) {
		warn "unbalanced RNA struct.\n";
	}
}
sub which_arm {
	my $substruct=shift;
	my $arm;
	if ($substruct=~/\(/ && $substruct=~/\)/) {
		$arm="-";
	}
	elsif ($substruct=~/\(/) {
		$arm="5p";
	}
	else {
		$arm="3p";
	}
	return $arm;
}
sub biggest_bulge {
	my $struct=shift;
	my $bulge_size=0;
	my $max_bulge=0;
	while ($struct=~/(\.+)/g) {
		$bulge_size=length $1;
		if ($bulge_size > $max_bulge) {
			$max_bulge=$bulge_size;
		}
	}
	return $max_bulge;
}
sub get_asy {
	my($table,$a1,$a2)=@_;
	my ($pre_i,$pre_j);
	my $asymmetry=0;
	foreach my $i ($a1..$a2) {
		if (defined $table->{$i}) {
			my $j=$table->{$i};
			if (defined $pre_i && defined $pre_j) {
				my $diff=($i-$pre_i)+($j-$pre_j);
				$asymmetry += abs($diff);
			}
			$pre_i=$i;
			$pre_j=$j;
		}
	}
	return $asymmetry;
}

sub peaks{
	my @cluster=@{$_[0]};

	return if(@cluster<1);

	my $max=0; my $index=-1;
	for (my $i=0;$i<@cluster;$i++) {
		if($cluster[$i][2]>$max){
			$max=$cluster[$i][2];
			$index=$i;
		}
	}
#	&excise(\@cluster,$index,$_[1],$_[2]);
	return($index);
}

sub find_cites{
	my @tmp=@{$_[0]};
	my $i=&peaks(\@tmp);
	
	my $start=$tmp[$i][0];
	my $total=0; my $node5=0;
	for (my $j=0;$j<@tmp ;$j++) {
		$total+=$tmp[$j][2];
		$node5 +=$tmp[$j][2] if($tmp[$j][0]-$start<=2 && $tmp[$j][0]-$start>=-2);
	}
	push @{$cites{$_[1]}{$_[2]}},$tmp[$i] if($node5/$total>0.80 && $tmp[$i][2]/$node5>0.5);
}

sub newpos{
	my ($a,$b,$c,$d)=@_;
	my $s= $a>$c ? $c : $a;
	my $e=$b>$d ? $b : $d;
	return($s,$e);
}

sub rev{

    my($sequence)=@_;

    my $rev=reverse $sequence;   

    return $rev;
}

sub com{

    my($sequence)=@_;

    $sequence=~tr/acgtuACGTU/TGCAATGCAA/;   
 
    return $sequence;
}

sub revcom{

    my($sequence)=@_;

    my $revcom=rev(com($sequence));

    return $revcom;
}

sub find_strand{

    #A subroutine to find the strand, parsing different blast formats
    my($other)=@_;

    my $strand="+";

    if($other=~/-/){
	$strand="-";
    }

    if($other=~/minus/i){
	$strand="-";
    }

    return($strand);
}
sub usage{
print <<"USAGE";
Version $version
Usage:
$0 -map -g -d -f -o -s -e
options:
  -map input file# align result # bst. format
  -g input file # genome sequence fasta format
  -d <int>   Maximal space between miRNA and miRNA* (200)
  -f <int>   Flank sequence length of miRNA precursor (10)
  -o output file# percursor fasta file
  -s output file# precursor structure file
  -e <folat> Maximal free energy allowed for a miRNA precursor (-18 kcal/mol)

  -h help
USAGE
exit(1);
}