The development of cytoplasmic male sterility (CMS) in the mid 1950's led to the commercial exploitation of F1 hybrids in grain sorghum in the USA. While other discovered CMS systems are available, commercial sorghum hybrid seed production relies on the A1 CMS system. The frequent and uniform use of the A1 CMS systems along with the photoperiod sensitive nature of exotic sorghum germplasm has limited genetic gains. However, conversion of tall photoperiod sensitive sorghum genotypes to short photoperiod insensitive types has lately made a significant impact and accounts for many of the improvements in sorghum hybrid production. Germplasm diversity in a breeding program directly affects the potential of genetic gain through selection, and allows the plant breeder to make a classification of germplasm into heterotic groups, which provides powerful information for hybrid development. The appropriate use of this information is crucial for the identification and development of parental inbred lines, one of the most time consuming phases in sorghum hybrid development. Since performance of lines per se does not necessarily predict hybrid performance, single cross hybrid tests for heterosis can be conducted to improve the efficiency in hybrid breeding programs. The objectives of this study were: (1) to estimate the genetic diversity present among a set of 21 sorghum converted lines (2) to classify this set of lines based on genetic similarities estimated using AFLP markers (3) to estimate heterosis, general and specific combining ability for grain yield among the set of converted lines. Field as well as molecular data has already been collected. Preliminary analysis of genetic similarities (cluster analysis), general and specific combining abilities and heterosis estimates are in progress.