Department of Molecular Biotechnology
University of Washington
Seattle, Washington, 98195
Random subcloning strategies are commonly employed in both physical mapping and DNA sequencing. Mathematical modeling is useful in planning such strategies and in providing a reference for their evaluation. Current analytic models for random subcIoning were originally developed for mapping projects and are accurate at low redundancy (Lander and Waterman, 1988). However, many sequencing projects are executed with higher redundancies and are not accurately modeled with existing theory. Additionally, genome spanning maps necessitate high redundancy cloning. The use of the beta distribution to model fragment start sites generated from finite genomes bridges many of the difficulties encountered by previous models. The general utility of the beta distribution and fragment start site approach is illustrated here with a new mathematical model for random shotgun strategies. Comparison with computer simulations and experimental data shows that this model is accurate at all redundancies.