Helitrons are loosely structurally defined, as their 5' and 3' terminii are a handful of nucleotides long. We apply HelitronScanner, but caution that many of these elements are candidates, and should not be believed to be true helitron elements without further study.
The script run_helitron_scanner.sh identifies direct and reverse complement orientation helitrons, converts the HelitronScanner output to a tabular format, grabs the final 30 bp of the element for clustering, does the clustering, assigns families, and outputs a gff.
Note that to read in entire maize chromosomes (and keep coordinates safe), HelitronScanner requires a lot of memory (I gave it 300 Gb, but I may have been overdoing it).