3.1 Protein content of parents and Teosinte derived BC1F2:3 population
The protein content of parental lines and introgressed populations is presented in table 2. The protein content of the pollen parent, Teosinte (Zea mays ssp. parviglumis), was higher (19.67%) than that of the maize inbred line CML-451 (9.02%) which was used as a seed parent. One way Analysis of variance was performed with a null hypothesis of no significant difference between genotypes for mean protein content. The results of the investigation however yielded a highly significant difference among genotypes for protein content. ANOVA therefore indicates that there are adequate variability for protein content in the materials chosen for the analysis (Table 1). The protein content of introgressed lines varies substantially, ranging from 9.53 % to 16.49 % for MP51 and MP6 lines, respectively. Out of 126 lines, four lines (MP6, MP20, MP36, and MP97) had protein content in the range of 16.07- 16.49 %.whereas two lines namely MP56 and MP99 had protein content of 15.05 and 15.07.% respectively. Twenty three lines (MP5, MP11, MP12, MP17, MP18, MP24, MP27, MP28, MP31, MP35, MP46, MP58, MP63, MP79, MP80, MP81, MP104, MP107, MP111, MP113, MP119, MP121 and MP124) of the 126 varied in protein content from 14.02- to 14.85% The lines possessed protein content in the range of 13.01 .to 13.95.% were twenty five in number (MP16,MP23,MP29, MP30, MP39, MP40, MP47, MP49, MP52, MP55, MP59, MP65, MP77,MP78, MP82, MP83, MP84, MP86, MP88, MP89, MP94, MP101, MP109, MP118, MP125) whereas another set of thirty seven lines (MP2,MP3, MP4, MP7, MP14, MP19, MP22,MP26, MP33, MP41, MP48, MP50, MP53, MP54 , MP61,MP60, MP62, MP64, MP66, MP70, MP71, MP74, MP87, MP92, MP93,MP96, MP98, MP102, MP103, MP105, MP110, MP112, MP115, MP116, MP117, MP123 and MP126) had protein content in the range of 12.00-to12.95 %. Sixteen derived lines namely MP13,MP15,MP32,MP42, MP43, MP44, MP45, MP57, MP68, MP72, MP73, MP85, MP90, MP100, MP114 and MP120 possessed protein content of 11.22-to 11.96% , fourteen lines (MP1, MP9, MP10, MP25, MP34, MP37, MP38, MP67, MP69, MP76, MP91, MP95, MP106 and MP108) had protein content from 10.02 to 10.94 %. Five lines namely MP8, MP21, MP51, MP75 and MP122 showed protein content of 9.53.to 9.98 %. (Table: 2). All the BC1F2:3 possessed protein content higher than the seed parent CML-451 but lower than the pollen parent teosinte. The increase in protein content in derived lines over CML-451 were varied from minimum of 9.53% in MP51to a maximum of 16.49 % in MP6 Comparison of protein content of 126 lines with protein content of teosinte indicates that the derived possessed minimum of 9.53% protein to maximum of 16.49. % protein content of the pollen parent teosinte. The data across the 126 lines indicated an average increase in protein content of 41.42% over the seed parent CML-451 whereas in comparison to pollen parent, the derived lines showed only up to 16.49 % protein content of pollen parent teosinte.
3.2. Ear traits of parents and Teosinte derived BC1F2:3 population
The one sample t test was performed to analyze individual line data without replication because data was recorded on F2:3 single cobs and found significant differences (p=0.001) among genotypes for all traits (Table:3). The Zea may ssp. parviglumis had ear characteristics of 5-6cm ear length, 0.51cm ear width, two kernel rows ear, contains 5-8 kernels per ear and test weight of 3.64g (naked kernels) and 7.04g (with seed coat) (Fig1,Table2). Maize inbred line CML-451 had ear length of 14.85 cm, ear diameter of 3.41 cm, 14 kernel rows, 30 kernels per row and test weight of 28.56g. The data on individual ears of parents and introgressed populations are illustrated in Table 3. The introgressed population exhibited wide variation for all ear traits. Ear length ranged from 6.8 to 23.5cm for MP74 and MP115 lines, respectively. Ear width varied from 2.0cm for MP7 and MP12 lines to 3.5cm for MP123 line. Number of kernel rows per ear also varied significantly and ranged from 8 (MP12, MP21, MP22, MP24, MP29, MP52, MP56, MP60 and MP68) to 16 (MP108 and MP118) lines respectively. Number of kernels per row varied from 4 (MP17, MP20 and MP21) to 30 (MP93) representing wide variation for number of kernels per row among teosinte derived maize lines. The results also indicated genetic variation for kernel weight; teosinte kernels are small enclosed within a hard seed coat and maize kernels are naked and bold, while teosinte derived maize lines exhibited differences for test weight ranging from 12.36g (MP60) to 30.58 g (MP111). Apart from ear characteristics, introgressed lines were observed for kernel shape. Teosinte seeds are black in colour, with brown pointed kernels, whereas maize kernels are bold, flat, and bright yellow in colour (Fig2). The teosinte derived BC1F2:3 population had a variety of kernel forms and observed round, flat and pointed shaped kernels (Table 2). Because of the introgression of genetic material from teosinte into the maize background, an introgressed offspring’s kernels shape has been modified. Among 126 teosinte derived lines, 73 lines had round shaped kernels, 33 lines had flat shaped kernels and remaining 21 lines showed pointed kernel shape. The simple linear regression analysis was done by plotting the protein content of each line against the kernel shape of the respective lines. The results revealed a non-significant correlation between protein content and kernel shape, indicating that kernel shape is not an effective trait for the selection of higher protein-content lines (Fig:3). However, when lines were grouped into three based on the shape of kernels, the round-shaped kernel group has a mean protein content of 12.73%, the flat-shaped kernel group has a 12.07% mean protein content, but the pointed kernel shape group has the highest mean protein content of 13.71 compared to other groups. Still, one notable observation was that all pointed kernel-shaped lines contained more than 12% protein content, and MP97 introgressed line displayed a pointed kernel shape with a protein content of 16.28%. Hence, the data disclosed that the shape of kernels was not enough to select the lines having higher protein due to a lack of clear distinction of introgressed lines into maximum and minimum protein content lines based on kernel shape.
Principal component analysis:
The variability amongst the Teosinte derived BC1F2:3 lines was analyzed using principal component analysis (PCA). Out of the six PCs, the first three explained 75.61 percent of the total variation in the data (Table 4). The perusal of the Table 4 indicated that the first, second, third, fourth, fifth, and sixth PCs accounted for 41.81, 20.47, 13.33, 12.11, 8.27, and 4.01 percent of the total variation, respectively. The PC1 accounts for 41.81 percent of total variability with positive eigen values for Ear Length (0.3634), Ear Width (0.5562), No of Kernel Rows/Ear (0.4253), No. of Kernels per Row (0.4418) and test weight (0.1853) while eigen value for protein content on first axis was negative (-0.3896). Ear Width was the character which accounted for maximum variability on PC 1. The PC2 accounted for 20.47% of total variance of the data The major characters that made a significant contribution to the second component were protein content (0.1232 ), Ear Length (0.4833 ), Ear Width(0.0098) and test weight (0.7349) which possessed positive Eigen values whereas No. of. Kernels/Row (-0.1900), No of Kernel Rows/Ear (-0.4183) had negative Eigen values on second principal component. The PC3 had 13.33 % of the total variability observed in population and traits positively contributed to this axis were Ear Width (0.1545), No of Kernel Rows/Ear (0.4106) and test weight (0.4501) while protein content (-0.2224), Ear Length (-0.4946 ) and No. of. Kernels/Row (-0.5576) showed negative Eigen values on PC3. However, test weight trait, provided more variability to PC2 and PC3, with values of 0.7349 and 0.4501, respectively.
3.4. Cluster analysis
One hundred twenty-six Teosinte derived maize lines along with parents were subjected to cluster analysis using Ward’s method agglomerative clustering , applying squared Euclidean distance as the distance measure, and were grouped into 6 clusters based on protein content and test weight of kernels (Table 6). The cluster I was consisted of Teosinte (Zea mays ssp. parviglumis) which had protein content of 19.67% and test weight of 3.645g. Cluster II grouped 4 lines (MP46, MP97, MP99 and MP111) in which mean of protein content was 15.13% and kernels test weight of 27.32g. Clusters III had 28 lines (MP6, MP11, MP12, MP17, MP20, MP24, MP27, MP31, MP35, MP36, MP49, MP52, MP56, MP63, MP65, MP77, MP78, MP79, MP80, MP83, MP84, MP86, MP88, MP101, MP104, MP113, MP119 andMP124,) having 14.25% and 20.47 g cluster means for protein content and test weight, respectively. Cluster IV possessed maximum number of lines (46) with mean protein content of 12% and a test weight of 23.51g. The seed parent, CML-451, was also grouped in this cluster. Thirteen lines grouped in Cluster V (MP34, MP37, MP38, MP44, MP69, MP75, MP76, MP90, MP91, MP95, MP108, MP114 and MP122) has mean protein content and test weight of 10.60 % and16.14g respectively. Cluster VI contained 36 lines (MP5, MP13, MP16, MP18, MP26, MP28, MP39, MP40, MP42, MP43, MP45, MP47, MP53, MP54, MP57, MP58, MP60, MP61, MP71, MP72, MP74, MP81, MP82, MP87, MP89, MP94, MP96, MP98, MP100, MP103, MP107, MP110, MP120, MP121, MP125 and MP126) in which cluster mean for test weight was 16.12g protein content of 12.92%. The dendrogram was constructed by the Ward method based on Euclidean distance coefficient matrices, categorizing parents and 126 teosinte introgressed lines into 6 major clusters based upon the Euclidean distance of protein content and test weight of kernels combined together (Fig: 4). The maximum diversity was observed between ClusterI (Teosinte) and the rest of the clusters with a Euclidean genetic distance of more than 8. This classification re-affirmed the greater genetic variability among teosinte-derived maize lines for test weight and protein content of kernels. Teosinte falls under one cluster which might be due to having entirely different test weight and protein content than maize-derived lines. Within each cluster, more similar lines were grouped and genetically dissimilar lines were grouped between clusters. Forming larger number of clusters represents the wider diversity for traits considered for classification of lines and data represented in the Table 6.The scatter plot matrix (Fig: 5) was also created to know the association between protein content and test weight of kernels and observed an interrelation of protein content and test weight of kernels but in a negative direction.
Correlation analysis: A correlation study indicated a different degree of association among the ear traits and protein content of teosinte derived lines and parents. Protein content exhibited significant negative correlation coefficients of -0.1911, -0.3861, -0.3249, and -0.3391 and non-significant correlation coefficient of -0.1036 with ear length, ear width, number of kernel rows per ear, number of kernels per row and test weight of kernels, respectively (Table: 5) Ear length was significantly and positively correlated with ear width, number of kernel/rows per row and test weight of kernels. Similarly, ear width was significantly and positively correlated with number of kernel rows per ear, number of kernels per row and test weight. Number of kernel rows per ear exerted significant positive association with number of kernels per row while non-significant negative with test weight. Number of kernels per row implied negative correlation with test weight.