Re^12: Help with pushing into a hash

[jemswira]

Hey there kenosis. Sorry to bother you again. SO the program is no longer giving me error messages, but the pushing to the arrays here:

  push @activline, "$1 | $2 | $activ{$1} \n" if $activ{$1};
    push @antioxline, "$1 | $2 | $antiox{$1} \n" if $antiox{$1};
    push @toxinline, "$1 | $2 | $toxin{$1} \n" if $toxin{$1};

isn't working properly. For example I selected a specific number from $activin, which was in $uniprot, but it wasn't in the final output. So @activline, @antioxline and @toxinline are much smaller than they should be. So a lot of the data that is supposed to be in the output is not there anymore. I tried testing with this:

if ($1 eq "E7F888"){print $2;}

and it gave me the correct $2, I tried this too:

next unless /(E7F888)\s+.+=([^\s]+)/;
 push @activline, "$1 | $2 | $activ{$1} \n" if $activ{$1};

print @activline;

and it printed the correct thing. But if I removed the tests i put, and ran the normal code, E7F888 is not in @activline. Is there anything I could do to figure out what's wrong with my code? Here's it again.

#!/usr/bin/perl
use Modern::Perl;
use File::Slurp qw/read_file write_file/;

my $uniprot     = 'uniprot-ACN';
my $activin    = 'Activator-PFAM.txt';
my $antioxin = 'AntiOxidant-PFAM.txt';
my $toxinin= 'Toxin-PFAM.txt';




my $activout = 'ActivACNPF.txt';
my $antioxout= 'AntioxACNPF.txt';
my $toxinout= 'ToxinACNPF.txt';
my @activline;
my @antioxline;
my @toxinline;

my %activ  = map {s/\.\d+//g; /(.+)\s+\|\s+(.+)/ and $1 => $2 } grep /
+\|\s+\S+/, read_file $activin;
my %antiox = map { s/\.\d+//g; /(.+)\s+\|\s+(.+)/ and $1=>$2; } grep/\
+|\s+\S+/,read_file $antioxin;
my %toxin  = map { s/\.\d+//g; /(.+)\s+\|\s+(.+)/ and $1=>$2; } grep/\
+|\s+\S+/,read_file $toxinin;
#if (exists $activ{"E7F888"}) {
#print $activ{'E7F888'};}
#else {print "LOL";}
for ( read_file $uniprot ) {
 next unless /(.{6})\s+.+=([^\s]+)/;
   # if ($1 eq "E7F888"){print $2;}
    
    push @activline, "$1 | $2 | $activ{$1} \n" if $activ{$1};
    push @antioxline, "$1 | $2 | $antiox{$1} \n" if $antiox{$1};
    push @toxinline, "$1 | $2 | $toxin{$1}  \n" if $toxin{$1};
 
}
print @activline;
write_file $activout, @activline;
write_file $antioxout, @antioxline;
write_file $toxinout, @toxinline;

THanks so much.

[Kenosis]

Hi, jemswira!

This is no bother at all!

You mentioned that the following captured a key that wasn't in the final set:

next unless /(E7F888)\s+.+=([^\s]+)/;
 push @activline, "$1 | $2 | $activ{$1} \n" if $activ{$1};

print @activline;

Try the following:

next unless /(E7F888)\s+.+=([^\s]+)/;
say;
exit;

saywill print the scalar $_ which the regex operates on. By doing this, you can examine the line to verify that the matches the following regex:

/(.{6})\s+.+=([^\s]+)/

Also, it may be that the 'key' is not being captured from $activin, since:

push @activline, "$1 | $2 | $activ{$1} \n" if $activ{$1};

will work only if the 'key' if found in both files and the regexs which process both work as they should.

Let me know what you find...

[jemswira]

well i ran the code and it returned this:

E7F888 Name=arid5b ;PF01388

which is correct. Well I selected the number cause it was in both files ( I manually ctrl-f ed the notepad file.

Well a lot of lines are missing from the final output. The $activ is 512 kb large and the out put is 3kb. And also @activline is missing alot when I did print @activline;. I randomly selected E7F888 from $activin and checked to make sure it was in $uniprot, so it may be another offending line of data. If I did this:

print $activ{E7F888},"\n";
for ( read_file $uniprot ) {
  next unless /(E7F888)\s+.+=([^\s]+)/;
say;    
    push @activline, "$1 | $2 | $activ{$1} \n" if $activ{$1};

}
print @activline;
  
}
print @activline;

I got this:

PF01388
E7F888 Name=arid5b ;PF01388

E7F888 | arid5b | PF01388

which showed that it was in %activ and that it was found in the file. But when I run through the normal code, it is not in @activ anymore.

UPDATE,

I ran the following code to check something, and not all the data is printed, I.E. A lot of lines are not pushed into the array.

#!/usr/bin/perl
use Modern::Perl;
use File::Slurp qw/read_file write_file/;

my $uniprot     = 'uniprot-sfinal.txt';
my $activin = 'Activator-PFAM.txt';
my %activ = map { s/\.\d+//g; /(.+)\s+\|\s+(.+)/ and $1=>$2; } grep/\|
+\s+\S+/,read_file $activin;
for ( read_file $uniprot ) {
    next unless /(.{6})\s+.+=([^\s]+)/;
    print $1,"\n"
     if $activ{$1};
}
print "done";

I am sure that there are more lines that are supposed to be printed. E.G.

$uniprot:

Q801F8  Q90XZ5; Q90XZ8;Name=dmrt1 
B7ZS42  A4PBN7; B7ZS44;Name=dmrt1-b ;PF00751;PF12374
Q157S1 Name=Crtc1  Synonyms=Mect1, 
$activ:
Q801F8 | PF00751.13 PF12374.3 PF12374.3
B7ZS42 | PF00751.13 PF12374.3 PF12374.3
Q157S1 | PF12886.2 PF12885.2 PF12884.2

I shall try to reverse the order, I.E, make a hash out of $uniprot instead of $activ and run it again.

[Kenosis]

Perhaps your data is more complex than first thought. You listed the following:

$uniprot
Q801F8  Q90XZ5; Q90XZ8;Name=dmrt1 
B7ZS42  A4PBN7; B7ZS44;Name=dmrt1-b ;PF00751;PF12374
Q157S1 Name=Crtc1  Synonyms=Mect1, 
$activ:
Q801F8 | PF00751.13 PF12374.3 PF12374.3
B7ZS42 | PF00751.13 PF12374.3 PF12374.3
Q157S1 | PF12886.2 PF12885.2 PF12884.2

Give the above listing, are you (potentially) expecting the following (keys/values)?

Q801F8 -> PF00751 PF12374 PF12374
B7ZS42 -> PF00751 PF12374 PF12374
Q157S1 -> PF12886 PF12885 PF12884

...

Q801F8 -> dmrt1 PF00751 PF12374 PF12374
Q90XZ5 -> dmrt1 PF00751 PF12374 PF12374
Q90XZ8 -> dmrt1 PF00751 PF12374 PF12374
B7ZS42 -> dmrt1-b PF00751 PF12374 PF12374
A4PBN7 -> dmrt1-b PF00751 PF12374 PF12374
B7ZS44 -> dmrt1-b PF00751 PF12374 PF12374
Q157S1 -> Crtc1 PF12886 PF12885 PF12884

It looks like there are multiple keys on a single $uniprot line:

Q801F8  Q90XZ5; Q90XZ8;Name=dmrt1
^^^^^^  ^^^^^^  ^^^^^^
   |       |       |
   +-------+-------+--- keys to be captured?

If this is the case, the current regex operating on the $uniprot lines would fail.

Let me know if the script needs to capture potentially multiple keys on a single $uniprot line...