untitled

232 10 Transforming with Traditional Programming Languages

while (<>) {

chomp;

if (/^MATRIX ([0-9]+)$/) {

$label = $1;

} elsif (/^number of sequences = ([0-9]+)$/) {

$numberOfSequences = $1 + 0;

} elsif (/^[ACGT] [|]/) {

@record = split;

for ($i = 2; $i < scalar(@record); $i++) {

$motifs{$label}{$i-2}{$record[0]} =

$record[$i] / $numberOfSequences;

}

}

}

foreach $label (sort(keys(%motifs))) {

print "Probability distributions for motif $label\n";

%motif = %{ $motifs{$label} };

foreach $position (sort(keys(%motif))) {

foreach $base (A, C, T, G) {

print("$base $motif{$position}{$base} ");

}

print("\n");

}

}

Program 10.13 Extracting data structures from a file using pattern matching

the first DNA base is the third field on the line, the second frequency is the

fourth field, and so on. So it is necessary to subtract 2 from the field position

to get the DNA base position. Finally, an item in the third dimension is the

probability for one of the four DNA bases. This is obtained by dividing the

frequency by the number of sequences.

Having extracted the motifs, the next step is to print them. Since the motifs

are in a 3D data structure, the most natural way to use the structure is with

three nested loops. The first loop processes the motifs. The labels are the

keys of themotifshash, and it is customary to sort the keys of a hash so

that they are printed in a reasonable order.