note ww SuperSearch (done already - here's the link: [href://?node_id=3989;BIT=FASTA] -- will give you a short list of recent discussions on dealing with FASTA files. <p>My notion that your paragraphing might be identifiable with a regex is pretty useless here. However, there's no reason you can't read a 2 lines at a time and use hashes to ensure the two "neighbors" values are discrete.</p> That too, however, breaks down if the dups appear other than adjacent to one another, given the size of your data. <p>So if none of the above help, you may wish to read about bioperl at both the wikipedia article, [http://en.wikipedia.org/wiki/BioPerl] and at the project page, [http://www.bioperl.org/wiki/Main_Page].</p> 1007560 1007807