"I would like to pull out all of the full length CDS with some 5' and 3' UTR info for the proteins I have found in the tissue by mass spec"

If I understand you correctly, I will look at the gff/gtf coordinates for the CDs feature and try to obtain that for the specific protein annotation in the gff file, A fastA file may not have these features handy so it maybe not the way to go unless you have an annotated genome somewhere for you to extract this stuff from.

• intro to GFF format
• GFF Object modules
• extension of GFF object modules

"I have tried simply blasting the peptides against the assembly but I get hits that are not in open reading frames"

Maybe because BLAST is a local alignment and search tool, that behavior is totally expected.

" I would suppose the major difference is that I need to return all the ORFs from tens of thousands of entries as a new fasta file."

Read the file one line at a time. Maybe something like what follows ?... If you don't have BioPerl installed then certainly there are many other examples on reading fastA file that don't use BioPerl around here

#UNTESTED CODE TO DEMONSTRATE A WAY AROUND
use strict;
use warnings;
use Bio::SeqIO;

my $in = Bio::SeqIO->new(-file =>"FastA.fa", format = "FASTA");
while(my $seq = $in->next_seq){
    my $sequence = $seq->seq;
    #Do something with $sequence
    my $ORFpattern = "foo";
    if($sequence =~ /$ORFpattern/){
        #report or anything
        }
    }
[download]

"Hi Pearl Monks"

Welcome to the Monastery and good luck....