This repository contains precomputed ROCker models. By using these models, you won't need to simulate datasets, just do a similarity search (BLASTx) using the .ref.fasta as a database and then filter the results using a ROCker model that adjusts to the read length of the query short-read metagenome.