BERGAMO, ITALY
Istituto Sperimentale per la Cerealicoltura
Genetic relatedness and variability among maize inbred lines selected in Italy --Hartings, H, Chittò, A, Bertolini, M, Verderio, A, Motto, M In breeding programs, information on genetic relationships within and between species is used for organizing germplasm collections, identifying heterotic groups within crops, and selecting parents for purposes of crossing. In this respect, DNA fingerprinting offers the possibility of studying genetic variation and relationships at the molecular level.

To address the issue of genetic relatedness and variability between the inbred lines developed in Italy by the Bergamo Maize Station, an AFLP analysis (Vos, P et al., Ac. Res. 23:4407-4414, 1995) was performed on a series of 71 inbred lines, considered representative of optimized breeding material. Twenty reference lines, encompassing the major heterotic groups available in the U.S. Corn Belt were included in the analyses in order to maximize genetic variability across the data set. Hence, the reference lines supplied a basis of genetic diversity to which the Italian inbred lines were related in the evaluation of their relative genetic relationships.

AFLP analysis of the Italian and reference inbred lines produced stable and repeatable profiles, which allowed us to unequivocally fingerprint each inbred line analyzed. A total of 682 polymorphic bands were revealed by the use of 5 PstI/MseI and 9 EcoRI/MseI primer combinations (PCs). E/M PCs displayed between 29 and 65 polymorphic bands resulting in an average of 47.56 12.68 markers. P/M PCs appeared less variable, disclosing between 41 and 54 polymorphic bands with an average of 50.8 5.63 markers. E/M PCs displayed an average Polymorphism Information Content (PICav) value of 0.33 0.027, while an average PICav of 0.35 0.014 was obtained with P/M PCs. In general, PICav values ranged from 0.28 to 0.36, demonstrating the good discriminatory power of the markers identified.

Scoring of the markers allowed the construction of a 682x91 binary array, which was consequently utilized to compute genetic distance (GD) values (Nei, M and Li, WH, Proc. Natl. Acad. Sci. USA 76:5269-5273, 1979) for all pairs of inbred lines considered. GD values ranged from 0.124 for inbred lines Lo876 and Lo1064, both derived from Lo876o2, to 0.62 for inbreds Lo3 and Lo903. An average GD of 0.437 0.012 was calculated for the entire data set.

Distance measures were subsequently used to construct a hierarchical tree using the UPGMA method using an NTSYS-PC program. Table 1 summarizes the grouping of inbred lines, presenting the major groups identified as well as their disclosed subgroups. Cluster analysis largely agreed with pedigree information as can be established by comparing the pedigree with clustering information.

Table 1. Principal groups of inbred lines as identified by cluster analysis.
 
 
BSSS1 (22)a
BSSS2 (9) References (12)
  Lo950 Lo903 Lo999 Lo876 Lo986
  Lo951  Lo904 Lo1055 Lo1016 A71
  Lo960 Lo1086 Lo1137 Lo1064 A69Y
  Lo964 Lo1101 Lo1141 Lo1066 B57
  Lo1054 Lo1106 B37 Lo1067 CI187-2
  Lo1053 Lo1127   Lo1123 FR5
  Lo1087 Lo1167   Lo1169 H55
  Lo1094 B73   Lo1170 N6
  Lo1173      A632 Oh07
          Os420
          W64A
          Wf9
Aveb   0.32   0.37 0.47
Min   0.13   0.12 0.27
Max   0.48   0.49 0.54
           
 
LSC1 (21)
LSC2 (18)
Lo932 (4)
  Lo1059 Lo863 Lo881 Lo902 Lo932
  Lo1061 Lo1077 Lo1038 Lo924 Lo937
  Lo1063 Lo1095 Lo1056 Lo976 Lo944
  Lo1076 Lo1096 Lo1035 Lo1124 W153
  Lo1156 Lo1125 Lo1090 Lo1126  
  Lo1157 Lo1128 Lo1140 Lo1142  
  Lo1158 Lo1172 C103 Lo1166  
  Lo1159 Lo1176 Va59 Mo17  
  Lo1160 Lo1182 T8 Oh43  
  Lo1162        
  Lo1168        
  Lo1171        
Ave
0.37
0.40
0.33
Min
0.13
0.15
0.19
Max
0.51
0.54
0.47

anumbers in parentheses represent the number of inbreds in each group;

bAve: average GD, Min: lowest GD, Max: highest GD within groups.

BSSS: Iowa Stiff Stalk Synthetic; LSC: Lancaster Sure Crop

For data with a hierarchical structure, analysis of molecular variance (AMOVA) allows the study of patterns of genetic variation within and between groups through the examination of variance. This assay can be extended to evaluate molecular marker data even in the absence of replicated values for samples (Law, JR et al., Euphytica 102:335-342, 1998). An AMOVA of the AFLP data based on the grouping obtained in cluster analysis of the inbred lines considered is presented in Table 2. Clusters were used to recompose, in broad terms, BSSS, LSC, and unrelated heterotic groups. The amalgamation into heterotic groups was performed using both a small number of larger clusters, as well as a larger number of clusters of reduced size. In both cases, the within-population (clusters) components of variance dominated the AMOVA, accounting for 73% to 79% of the variation, with less than five percent representing variation between heterotic groups. Changes in the grouping pattern applied had no significant effect on the distribution of variation. Furthermore, the genetic distance between clusters (Fst values) exceeded both the degree of inbreeding within clusters (Fsc values) and the degree of relatedness between genes within inbred lines (Fct values) in all cases.

Table 2. Summary of AMOVA for AFLP data from Italian and U.S. Corn Belt Inbred Lines.
 
Nested AMOVA Variance Components
  Between Groups Between Populations Within Groups Within Populations      
Populations V(A)% V(B)% V(C)% Fst Fsc Fct
Large 2.61 18.31 79.1 0.209 0.188 0.026
Small 4.27 22.38 73.4 0.266 0.234 0.043

Heterogeneity within breeding groups was further analyzed by computing a series of diversity statistical indices from the AFLP data (Table 3). Estimates of q, which is the product of population gene number and mutation rate, were computed based on the number of polymorphic sites (ˆqs) and on the mean number of pairwise differences (ˆqp). Both estimates hold under the assumption of random mating, population equilibrium and neutral mutations. If these assumptions are valid, ˆqs has smaller stochastic variance than ˆqp. However, since ˆqp is independent of sample size while ˆqs is not, ˆqp may be a more reliable estimate of gene diversity within heterotic groups. None of the breeding groups analyzed showed significant variation in ˆqs and ˆqp values. Subsequently, the average gene diversity per site was computed. This parameter of diversity, as determined on the entire population of inbreds analyzed, equaled 0.33 ± 0.16 and varied within a narrow range (0.24 — 0.33) when determined for the heterotic groups identified. Expansion or contraction of the heterotic groups was assayed by statistics developed by Tajima (Tajima, F, Genetics 123:585-595, 1989). Tajima’s D statistics, computed on the heterotic groups listed in Table 3, did not reveal any expansion or contraction of groups, since none of the values obtained reached statistical significance using coalescent simulation (Hudson, R.R., Oxford Surways in Evalutionary Biology, Oxford Univ. Press, 1-44, 1990), or parametric approximation assuming a beta-distribution (Tajima, 1989).

Table 3. Summary of diversity statistics for Italian and U.S. Corn Belt Inbred Lines.
 
   
Groups
 
All Inbreds
Referencea
BSSS1
BSS2
LSC1
LSC2
No. inbreds
91
20
22
9
21
18
No. polymorphic sites
682
475
442
412
235
329
Theta(S)
0.16
0.33
0.27
0.36
0.28
0.29
S.D.b Theta(S)
0.04
0.13
0.09
0.15
0.09
0.10
Theta(p)
0.25
0.38
0.36
0.40
0.36
0.35
S.D. Theta(p)
0.12
0.20
0.17
0.22
0.18
0.18
Average gene diversity
0.33
0.33
0.24
0.26
0.27
0.30
+/- average diversity
0.16
0.17
0.12
0.14
0.14
0.15
Tajima’s D
2.24
0.68
0.92
0.45
1.15
0.91

aAll Reference lines; bS.D. Standard Deviation

In this study, AMOVA, performed on genetic structures obtained by means of two different amalgamation schemes showed that the within cluster component of variance largely exceeds the variance between heterotic groups, regardless of the amalgamation scheme selected. Moreover, additional statistical indices of diversity show that no significant differences occur between the variability encountered within identified heterotic groups and the overall level of genetic diversity among the inbred lines taken into consideration. It can therefore be concluded that breeding activity has by no means caused a decline of genetic variability within heterotic groups. On the contrary, levels of genetic diversity have remained substantially unchanged over time and hence, plant breeding has resulted in a qualitative rather than a quantitative shift in diversity. In conclusion, this study has shown that a large genetic variability occurs among maize germplasm available in Italy. This variability can be exploited in hybrid and line development for further yield improvement.
 
 


Please Note: Notes submitted to the Maize Genetics Cooperation Newsletter may be cited only with consent of the authors.

Return to the MNL 76 On-Line Index
Return to the Maize Newsletter Index
Return to the Maize Genome Database Page