User's Guide to the Program NewHybrids Version 1.0
Eric C. Anderson · 2002
NewHybrids is a program for computing the posterior distribution that individuals in a sample fall into difierent hybrid categories. The details of the algorithm are published in Anderson and Thompson (2002). This document describes how to use the program. In order to make the program accessible to researchers, and also for my own use of the program, I have created a graphical interface that allows the observation of most of the variables involved in the Markov chain Monte Carlo (MCMC) simulation. This is valuable for assessing the reliability of the results from the MCMC run. In order to take full advantage of these features, you will want to read A User’s Guide to the GLUT for Markov Chain Monte Carlo Graphical Interface which is included in the standard distribution of the NewHybrids software as the flle GFMCMC UserGuide.pdf. If you are like me, you will flnd it nice to start using the software immediately. For this reason, I have included a test data flle and describe how to get the program running on that as quickly as possible in the \Quick Start section (page 2). After that I go into more detail on the required data flle format, program options, and the graphical features unique to NewHybrids. This version (1.0) is the flrst distributed version. It lacks some features that I hope to add to the software very soon. In my mind, the greatest deflciency is that I have not added any features whereby the priors can be appropriately manipulated, or where samples of known individuals can be used to generate an appropriate prior. A further shortcoming is that I am not yet retaining (in a user-accessible way) the codes given to difierent alleles in the data set. Within the program, alleles get numbered internally from 0 to the the number of alleles at the locus minus one, but that internal numbering system may correspond in a mysterious fashion to how the alleles are numbered in your dataset. I will address these issues in a later version. There are bound to be some bugs in the program. I have developed it under the Mac Operating System, but have ported it to Windows. Since I don’t have a real Windows machine at my disposal, I have not been able to test the program extensively in that environment. Please be patient! And don’t hesitate to send me email, or call me at my o‐ce: (510) 643-6299.