MetaGenomeThreader: A Software Tool for Predicting Genes in DNA-Sequences of Metagenome Projects

We consider a gene finding method that is specifically designed to work on metagenome sequences. The method can handle short metagenome sequences with in-frame stop codons as well as frame shifts. It delivers gene predictions for a set of metagenome sequences, which may be individual reads or a collection of assembled reads sequenced from an environmental sample. The method searches for stretches of DNA that are conserved within the environmental sample. Conserved coding sequences are discriminated from conserved non-coding regions based on their synonymous substitution rate. We describe the program MetaGenomeThreader which implements the method and show its application on a synthetic metagenome.