Generalized Gene Cluster Detection Using CLOCI

Metabolic gene clusters (MGCs) are genomic loci that contain multiple genes that are functionally and genetically linked. MGCs collectively encode a spectrum of metabolic functions, including small molecule biosynthesis, nutrient assimilation, metabolite degradation, and production of proteins essential for growth and development. Due to their diverse ecological functions, identifying gene clusters is a powerful tool for small molecule discovery and provides insight into the ecology and evolution of organisms. Gene cluster detection algorithms have historically been specialized for detecting biosynthetic gene clusters that contain canonical "core" biosynthetic functions, while overlooking uncommon or unknown cluster classes. These overlooked clusters are a potential source of novel natural products and comprise an untold portion of overall gene cluster repertoires. Unbiased, function-agnostic detection algorithms therefore provide an opportunity to reveal novel classes of gene clusters and more precisely define genome organization.We developed CLOCI (Co-occurrence Locus and Orthologous Cluster Identifier) as a generalized, unbiased gene cluster detection algorithm. CLOCI generalizes gene cluster detection by identifying signatures of coordinated gene evolution that underlie all classes of MGCs. CLOCI first detects selection on gene colocalization by identifying and circumscribing shared synteny loci across a dataset of genomes into homologous locus groups. Gene clusters comprise a subset of these homologous locus groups, and CLOCI implements orthogonal proxies of coordinated gene evolution, such as quantifying loss and horizontal transfer of a locus, to enrich MGCs from homologous loci. Here, we describe the conceptual framework of the CLOCI algorithm and present a description of its implementation (see Note 1).