Re: how do i obtain blast result from the given file

You can print the matching lines with a regular expression that matches only the lines you're interested in:

    while (<$fh>) {
        print if /^(.+?)\|(.+?)\| (.+?)\s+(\d+)\s+(\d+e[+-]\d+)$/;
    }
[download]

The above will also handle arbitrarily large files, as it reads one line at a time into memory.

There are many hits, but if I want the top ten only, what do i do?

I'm not sure what you mean by top ten, exactly (highest "Score"? "E Value"? Order within the file?), but a likely approach to this would be to modify the above while loop and instead of printing, place each line in a hash. My regular expression already does basic parsing of the input line into capture variables $1 through $5, so you can try something like this:

    $hash{$2} = { col1 => $1, desc => $3, score => $4, E => $5 }
        if /.../; # Use regex above
[download]

Then, after the loop finishes, you can sort and display the results however you like. For example:

    my $how_many = 10;
    for (sort { $hash{$b}->{score} <=> $hash{$a}->{score} } keys %hash
+) {
        printf "%-4s %-55s %5s\n",
            $hash{$_}->{col1}, $hash{$_}->{desc}, $hash{$_}->{score};

        last if --$how_many == 0;
    }
[download]

Outputs:

emb  plasma membrane H+-ATPase [Oryza sativa Japonica...       213
gb   hypothetical protein OsI_09609 [Oryza sativa Indi...      213
ref  Os03g0100800 [Oryza sativa Japonica Group] >...           213
gb   ATPase 11, plasma membrane-type [Aegilops tauschii]       208
ref  PREDICTED: ATPase 7, plasma membrane-type is...           207
ref  H(\+)-transporting atpase plant/fungi plasma...           207
ref  PREDICTED: plasma membrane ATPase 1-like [Br...           207
ref  PREDICTED: ATPase 7, plasma membrane-type is...           207
ref  autoinhibited H+ ATPase [Populus trichocarpa...           206
ref  PREDICTED: plasma membrane ATPase 1-like iso...           205
[download]

Comment on Re: how do i obtain blast result from the given file Select or Download Code

Replies are listed 'Best First'.
Re^2: how do i obtain blast result from the given file by bingalee (Acolyte) on Jun 17, 2013 at 19:59 UTC
correction- I'm so sorry, I meant the first ten hits	[reply]
Re^3: how do i obtain blast result from the given file by rjt (Curate) on Jun 17, 2013 at 20:15 UTC
In that case, just modify the initial loop to stop after ten hits: `my $how_many = 10; while (<$fh>) { if (/^(.+?)\\|(.+?)\\| (.+?)\s+(\d+)\s+(\d+e[+-]\d+)$/) { print; last if --$how_many == 0; } }` [download] If that's all you want to do, you don't need the hash. Note that you can still format the output with printf: `printf "%-4s %-55s %5s\n", $1, $3, $4;` You could still use the hash, as well, if you need to do further processing on the results. In that case, to preserve the order, either add an index to the hash ref for use with sort, or, simpler, push each key to an array as you find them: `my @hits; # Keys in order my $how_many = 10; while (<$fh>) { if (/^(.+?)\\|(.+?)\\| (.+?)\s+(\d+)\s+(\d+e[+-]\d+)$/) { $hash{$2} = { col1 => $1, desc => $3, score => $4, E => $5, key => $2 }; push @hits, $2; last if --$how_many == 0; } } # Go through the first ten hits, in order for (map { $hash{$_} } @hits) { # $_ contains the hash ref for each record printf "%-5s %-55s %5s\n", $_->{col1}, $_->{desc}, $_->{score}; }` [download]	[reply] [d/l] [select]
Re^4: how do i obtain blast result from the given file by bingalee (Acolyte) on Jun 18, 2013 at 19:52 UTC
thank you :)	[reply]


Think about Loose Coupling
	PerlMonks