Today’s sorghum breeding programs integrate both traditional and molecular approaches. When breeding new hybrids, each program develops inbred lines for test-cross evaluation. To develop inbred lines, most but not all sorghum breeding programs use some form of pedigree selection [4]. Test-cross evaluation is used to assess combining ability. Good general combining ability (GCA) is needed in the first test-cross and subsequent testing identifies the exact combinations with commercial potential. Predicting hybrid performance by evaluating inbred lines themselves is not typically an efficient use of resources [65].

If the new lines are seed parents (i. e., they do not restore fertility to male sterile inducing cytoplasm), they must be sterilized for use as a seed parent. Sterilization involves back — crossing to introgress sterility-inducing cytoplasm, which results in a sterile version of the seed parent. The use of off-season nurseries speeds the pace but sterilization and creation of an A-line (male sterile) version of the B-line (fertile seed parent) requires additional time. Since the advent of hybrids, sorghum germplasm has been placed into heterotic groups based on fertility restoration. While this classification was of necessity, there is some basis as most seed parents were in the kafir group while pollinator parents were typically cau- datum and durra types. Since then breeding has evolved these into legitimate heterotic groups [73].

Sorghum genetics and genomics are well advanced. The sorghum genome has been sequenced [74], two high-density genetic maps and one cytogenetic map have been con­structed [75-77]. Additionally, many agronomically important genes have been cloned [78-80] and a transformation system has been developed [81]. Finally, there are many published quantitative trait locus (QTL) studies in sorghum, and much of this information is being used for genetic characterization of sorghum germplasm in breeding programs.