09.12.2012 Views

Principles of Plant Genetics and Breeding

Principles of Plant Genetics and Breeding

Principles of Plant Genetics and Breeding

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

152 CHAPTER 9<br />

Covariance =−2.45<br />

Correlation =−0.757<br />

Intercept = 59.34<br />

Slope =−1.087<br />

St<strong>and</strong>ard error = 0.332<br />

Student’s t-value = 3.273; probability = 0.010<br />

The results indicate a significant negative association<br />

between seed oil <strong>and</strong> protein content. The breeding<br />

implication is that as one selects for high seed oil, seed<br />

protein will decrease.<br />

Table 9.3 Data for oil <strong>and</strong> protein content <strong>of</strong> soybean<br />

seed.<br />

Oil content Protein content<br />

(%) (%)<br />

20.1 35.7<br />

21.2 35.1<br />

19.5 33.2<br />

18.3 40.6<br />

19.0 37.5<br />

21.3 36.1<br />

19.8 39.5<br />

22.6 34.8<br />

17.5 39.1<br />

19.9 40.2<br />

Mean 19.92 37.68<br />

Simple linear regression<br />

Unlike simple linear correlation, simple linear regression<br />

is a relationship between two variables that involves<br />

cause <strong>and</strong> effect. There is a dependent variable (Y ) <strong>and</strong><br />

an independent variable (X). For example, grain yield<br />

depends on seed size, number <strong>of</strong> seeds per pod, etc. The<br />

changes in the dependent variable (effect) are brought<br />

about by the changes in the independent variable<br />

(cause). Another way <strong>of</strong> looking at it is that regression is<br />

a study <strong>of</strong> the relationship between variables with the<br />

objective <strong>of</strong> identifying, estimating, <strong>and</strong> validating the<br />

relationship.<br />

Simple linear regression has the mathematical form <strong>of</strong><br />

the equation <strong>of</strong> a straight line:<br />

Y = a + bX<br />

where Y = dependent variable, X = independent variable,<br />

b = slope <strong>of</strong> the regression line, <strong>and</strong> a = intercept on<br />

the y axis.<br />

Table 9.4 Data on plant yield <strong>and</strong> maturity <strong>of</strong> soybean.<br />

Yield (bushels) Days to maturity<br />

44 138<br />

40 136<br />

38 125<br />

35 118<br />

33 115<br />

32 111<br />

30 110<br />

28 109<br />

24 98<br />

18 93<br />

Mean 32.2 115.3<br />

Figure 9.2 The linear regression line.<br />

The regression coefficient is calculated as:<br />

b = [N(∑XY ) − (∑X)(∑Y )]/N(∑X 2 ) − (∑X) 2<br />

The data in Table 9.4 represent the yield <strong>of</strong> soybean<br />

corresponding to various days to maturity <strong>of</strong> the crop.<br />

The results <strong>of</strong> a regression analysis are as follows:<br />

Covariance = 110.04<br />

Correlation = 0.976<br />

Intercept =−27.03<br />

Slope (b) = 0.514<br />

St<strong>and</strong>ard error = 0.040<br />

Student’s t-value = 12.794; probability = 0.000<br />

The prediction equation is hence (Figure 9.2):<br />

ˆY i = 127.03 + (0.514)X i<br />

Y = a + bX

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!