Just for completeness, a BioRuby solution to count Q + N in 80-residue moving window.

require 'rubygems'
require 'bio'

Bio::FlatFile.foreach("myseqs.fa") do |sequence|
  sequence.seq.split("").each_cons(80) do |window|
    puts window.count("Q") + window.count("N")
  end
end

Here, each sequence is split into an array and the each_cons() array method provides the moving window - it returns nil if the sequence length is < 80.

If you use numpy, it's easy and fast to take an 80bp window using convolve

The script below assumes you're giving it a protein fasta and prints out sequences that are "enriched", but if you need the exact position where the enrichment occurs within each sequence, you'll need to add a couple lines (lookup numpy.where). runs in a few seconds for a 4.5MB file.

from pyfasta import Fasta
import numpy as np

fasta = 'orf_trans.fasta'
window = 80
cutoff = 30

f = Fasta(fasta)
kern = np.ones((window,))

enriched = 0
for k in f.keys():
    seq = np.array(f[k], dtype='c')
    # get a boolean array with True where AA is one of 'Q', 'N'
    a = (seq == 'N') | (seq == 'Q')

    # take an 80bp moving window.
    mw = np.convolve(a, kern, mode='same')
    # if it doesn't have any windows > 30, skip it.
    if mw.max() < cutoff: continue
    enriched += 1
    # can also add something here to find the exact loc of the enrichement(s)
    print k.split()[0]

print "\n%s out of %i sequences are enriched" % (enriched , len(f.keys()))

Well, a very didactic python solution would be:

import re
seq = 'ACGGTCGIWYDEAJCFSHDKANNNNNQNQNQNQNQNDGASDGADJHDAGJDAQNNAQHKJHSHDDAAHHQNNQNQ'
score_by_position = {}
gln_or_asn_matcher = re.compile('[QN]')
step = 10 # replace it with 80 on the real seq
cutoff = 5 # replace it with 30

# note that this ignores all windows smaller than $step at the left and right of the sequence    
for pos in range(0, len(seq)-step):
    score = len(gln_or_asn_matcher.findall(seq[pos:pos+step]))
    score_by_position[pos] = score
    if score > cutoff:
        print 'possible enriched region at position %s; %s' % (pos, seq[pos:pos+step])

# requires matplotlib
from matplotlib.pylab import bar, show
bar(score_by_position.keys(), score_by_position.values())
xticks(range(len(seq)), seq)
show()

of course, there it should be other tools that do this and in a more secure way. you should test this script with other values before using it.

Here is my java solution:

File 'SearchGlnAsn.java':

import java.util.LinkedList;

public class SearchGlnAsn
    {
    public static void main(String[] args) throws Exception
        {
        int c;
        int pos=0;
        int count_N_Q=0;
        LinkedList<Boolean> window=new LinkedList<Boolean>();
        StringBuilder name=new StringBuilder();
        while((c=System.in.read())!=-1)
            {
            if(c=='>')
                {
                name.setLength(0);
                while((c=System.in.read())!='\n') name.append((char)c);
                pos=0;
                count_N_Q=0;
                window.clear();
                }
            else if(!Character.isWhitespace(c))
                {
                pos++;
                boolean isAA=(c=='N' || c=='Q' || c=='n' || c=='q');
                if(isAA) count_N_Q++;
                window.add(isAA);
                if(window.size()==80)
                    {
                    if(count_N_Q>30)
                        {
                        System.out.println(name+"\t"+count_N_Q+"\t"+(pos-80));
                        }
                    if(window.remove()) count_N_Q--;
                    }

                }
            }

        }

    }

Compilation:

javac SearchGlnAsn.java

Execution:

curl -s ftp://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/complete/uniprot_sprot.fasta.gz |\
gunzip -c |\
java SearchGlnAsn |\
awk -F '    ' '{if($2>75) print $0; }'

sp|Q156A1|ATX8_HUMAN Ataxin-8 OS=Homo sapiens GN=ATXN8 PE=1 SV=1    79  0
sp|Q54UG3|Y1095_DICDI UPF0746 protein DDB_G0281095 OS=Dictyostelium discoideum GN=DDB_G0281095 PE=3 SV=1    76  415

 curl "http://www.uniprot.org/uniprot/Q156A1.fasta"
>sp|Q156A1|ATX8_HUMAN Ataxin-8 OS=Homo sapiens GN=ATXN8 PE=1 SV=1
MQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQ
QQQQQQQQQQQQQQQQQQQQ