As a well known fact, organisms form larger and complex multimodular (composite or chimeric) and mostly multi-functional proteins through gene fusion of two or more individual genes which have independent evolution histories and functions. We call each of these components a module. The existence of multimodular proteins may improves the efficiency in gene regulation and in cellular functions, and thus may give the host organism advantages in adaptation to environments. Analysis of all gene fusions in present-day organisms should allow us to examine the patterns of gene fusion in context with cellular functions, to trace back the evolution processes from the ancient smaller and uni-functional proteins to the present-day larger and complex multi-functional proteins, and to estimate the minimal number of ancestor proteins that existed in the last common ancestor for all life on earth. Although many multimodular proteins have been experimentally known, identification of gene fusion events systematically at genome scale had not been possible until recently when large number of completed genome sequences have been becoming available. In addition, technical difficulties for such analysis also exist due to the complexity of this biological and evolutionary process. We report from this study a new strategy to computationally identify multimodular proteins using completed genome sequences and the results surveyed from 22 organisms with the data from over 40 organisms to be presented during the meeting. Additional information is contained in the original extended abstract.
