RELATIONSHIPS AMONG EARLY EUROPEAN MAIZE INBREDS .3. GENETIC DIVERSITY REVEALED WITH RAPD MARKERS AND COMPARISON WITH RFLP AND PEDIGREE DATA

Molecular markers such as random amplified polymorphic DNAs (RAPDs) or restriction fragmenth length polymorphisms (RFLPs) have been proposed for characterizing genetic relationships of maize (Zea mays L.) germp lasm. A set of 57 elite inbred lines representative of early-maturing European flint and dent heterotic groups were assayed for RAPDs using 54 primers. Our main objectives were to (i) investigate the amount of variation for RAPDs in these materials, (ii) examine the usefulness of RAPDs for assigning inbred lines to heterotic groups, and (iii) compa re the genetic similarity (GS) based on RAPD data with MALECOT'S coanc estry (f) based on pedigree data and with the genetic similarity (GS) based on RFLP data. Out of 54 primers, 31 were informative and yielde d 278 bands, 234 of which were polymorphic among the 57 lines. GS, cal culated from RAPD data as JACCARD's coefficient, ranged from 0.57 to 0 .80 between unrelated (f = 0) pairs of lines. All dent inbreds showed a smaller mean GS to lines from the other heterotic group than to line s from the same heterotic group. The same holds true for 16 of the 24 flint lines analyzed. Principal coordinate analysis of GS estimates re sulted in separate groupings of flint and dent lines. Correlations bet ween the coancestry coefficient f and GS were moderate for related (f > 0) pairs of flint (r = 0.42) and dent (r = 0.49) lines. Correlations of genetic similarities based on RFLP data (GS) with those based on RAPD data (GS), were moderate for flint (r = 0.43) and dent (r = 0.46) lines, when they were calculated for sets of unrelated pairs of lines (f = 0). Using sets of related (f > 0) pairs of lines, the correlatio ns between GS and GS were moderate for flint (r = 0.47) but weak for dent (r = 0.26) lines. We conclude that RAPDs based on a limited set o f random primers are less suitable for determining the degree of genet ic relatedness between maize inbred lines than RFLPs selected for a go od coverage of the maize genome.