SPECSTAT - PC SOFTWARE FOR STATISTICAL-DATA ANALYSIS IN THE FIELD OF POPULATION-GENETICS

This set of applied programs, SPECSTAT, has been written in Turbo Pasc al-5 and is adapted for an IBM compatible PC-XT/AT using MS-DOS. SPECS TAT is a small software package (approximately 300K) which has high ef ficacy when performing calculations. SPECSTAT provides some new comput ational opportunities not provided by the existing statistical package s for analyzing allozyme population genetic data. Furthermore, it is a ble to carry out some simple transformations of quantitative traits (Q T) and the genotypic data matrix. These transformations are convenient for the investigation of QT and allele frequencies and for the analys is of the correlations between QT and allozyme genotypic codes. From t he vectors of genotypic codes for each sample, the following parameter s are computed: i) the allele frequencies with standard errors (SE), i i) chi-squared values of goodness-of-fit for die observed and expected Hardy-Weinberg genotype frequencies and iii) H(o), H(s) and F(is) val ues. The H(t), F(it), F(st), D(st)' and F(st)', statistics, chi-square d values in the heterogeneity test of the predominant alleles, and G-s tatistics of all-allelic heterogeneity among samples with Williams' co rrection (Sokal & Rolf, 1981) are also available. Calculations are per formed to produce single-locus and combined matrices of Nei's genetic distances Dn and Dm (standard and minimal unbiased), and the matrix of combined Kalabushkin's similarity metrics, which maximize small diffe rences. Other computational opportunities are provided as well. Output data are simple ASCII files and are organized in a compact mode which can either be used directly for publication or used after minor chang es in further calculations.