Currently, many computer programs are available for estimating evolutionary distances and reconstructing phylogenetic trees from molecular data. However, most of them are written for specific methods and cannot be interconnected easily because of their inflexible input and output file formats . MEGA presents an interactive, user-friendly platform for estimating evolutionary distances, reconstructing phylogenetic trees, and computing basic statistical quantities that are of evolutionary interest. MEGA has been developed specifically for use on IBM and IBM-compatible personal computers.
MEGA is designed to facilitate extensive sequence data analysis from an evolutionary perspective using a single program package. At the same time, the overlap between the methods implemented in MEGA and those in other existing evolutionary analysis programs has been consciously avoided. This is reflected in the exclusion of the maximum likelihood method (PHYLIP) and in the absence of extensive options for the maximum parsimony method (PAUP and MacClade) . Limitations on the memory size and relatively slower speeds of desktop computers (and the presence of many commercial and non-commercial programs) prompted the decision not to include sequence alignment methods in MEGA.
In this manual, chapter 1 (Getting Started) explains the procedures for installing and ruming MEGA and provides information on obtaining technical support. In the following chapter (Input Data and Formats), various data and input file formats are discussed. This chapter also elaborates on in-memory data editing features available in MEGA.
In chapter 3 , a brief explanation of various statistical quantities that are useful for studying the evolutionary change of DNA and amino acid sequences is presented . Chapter 4 (Distance Estimation) describes most of the important statistical methods currently used for estimating evolutionary distances. This chapter also explains the handling of al alignment gaps and missing data in the computation of evolutionary distances . Chapter 5 (Phylogenetic Inference) presents descriptions of different tree-building methods and related issues.
Basic features of interactive user-interface, sequence data presentation, phylogenetic-tree editing, context-sensitive help, and text-file editing and browsing are discussed in chapter 6 (User-interface). Chapter 7 (Walk Through MEGA) presents a tutorial on using MEGA for data analysis. Chapter 8 (Command Reference) explains the use of menus and commands present in the user-interface. Descriptions of cryptic errors and some possible remedies are provided in chapter 9 (Error Messages).
There are five Appendices included in this manual. Appendix A provides a list of computational and editing functions available in MEGA, whereas Appendix B gives a,list of common questions concerning the use of MEGA and their answers. Appendices C, D, and E provide other information useful to the users of MEGA.
The MEGA project was initiated by the suggestion of M.N. However, the entire set of computer programs in MEGA was written by S.K. and K.T. S.K. is responsible for designing the layout of the programs and writing most parts of the programs, whereas K.T. is responsible for developing the tree-editing algorithm and major portions of the branch-and-bound and heuristic search algorithms for maximum parsimony. M.N. is responsible for choosing the statistical methods that are included in MEGA. He is also responsible for writing chapters 4 and 5 in this manual, though the algorithms for the maximum parsimony method in chapter 5 were developed by K.T. and S.K. The other chapters were written by S.K. and edited by M.N.
Many friends and colleagues provided encouragement and assistance in the development of MEGA and during the preparation of this manual . Test versions of MEGA were used in the classroom teaching by Masatoshi Nei and Austin Hughes at the Pennsylvania State University and by Marcy Uyenoyama at Duke University. These provided excellent opportunities to test the programs, and the students drew our attention to many inconsistencies in the user-interface. Initial drafts of certain chapters of this manual were critically read by Carla Hass, Tatsuya Ota, Naoko Takezaki, Marcy Uyenoyama, and Thomas Whittam. We also appreciate comments and help from Kyung-Eui Han, Blair Hedges, Thomas Lopez, Paul Mosquin, Arun Roychoudhury, Claudia Russo, Andrey Rzhetsky, Steve Schaeffer, and Tanya Sitnikova. We sincerely acknowledge everyone else who assisted in some form during this endeavor and whose name we fail to mention. The development of MEGA was partially supported by research grants from the National Institutes of Health and the National Science Foundation to M.N.