Modify SAM CIGAR string to soft-clip the last n bases

softclip_cigar.pl

my $cigar = "30M40D50M";
my $n_soft = 71;

my $out_cigar = "";

# there are already at least $n_soft soft-clipped bases at the end of the CIGAR
# then do nothing
if($cigar =~ m/(\d+)S$/ and $1 >= $n_soft) {
        $out_cigar = $cigar;
}
else {  
        # incrementally fetch the last part of the CIGAR and sum up their lengths until $n_soft is reached
        my $sum = 0;
        while(1) {
                if($cigar =~ m/((\d+[SMINDH])*)(\d+)([NDH])$/) { # hard clips, dels and ref skips are not counted
                        $cigar = $1;
                }
                elsif($cigar =~ m/((\d+[SMINDH])*)(\d+)([SMI])$/) {
                        $sum += $3;
                        if($sum > $n_soft) {
                                $out_cigar =  $1 . ($sum - $n_soft) . $4 . $n_soft . "S";
                                last;
                        }
                        else {  
                                $cigar = $1;
                        }
                }
                else { # last field was read, but number of bases in alignment is < $n_soft
                        $out_cigar = $sum . "S";
                        last;
                }
        }
}

print "$out_cigar\n";