Dictionary-based methods and their applications in biology and medicine

Abstract

This study proposes methods to explore genome organization and identify genome interactions that do not rely on annotations and aim to work on whole genome data. These methods use string matching between collections of dictionaries that depict genomes with different levels of resolution. Each dictionary represents a mapping of the complete genome data into a set of unique fixed-length segments. The methods are inspired by biological mechanisms including restriction-modification systems and CRISPR-Cas defenses that use exact matching. The use of this string-oriented approach might help researchers better understand biological mechanisms and avoid many of the drawbacks associated with annotations. These methods shift the computational paradigm from looking for specific instances such as genes and other elements within a genome to "full-search" analysis without preconceived targets. We hypothesize that the development of efficient dictionary-based screening methods will lead to a better understanding of genome organization and genome interactions. The results of this study indicate that these methods can capture many biologically significant relationships not easily captured by traditional approaches. The results of this study contribute to (a) changing a computational paradigm for processing genome data; (b) developing new methods for analyzing genome organization and relationships between genomes; and, (c) identifying and evaluating potential genome interactions at a broader scale for biological and medical applications.