Maize Genetics Cooperation Newsletter vol 85 2009

 

RIO CUARTO, ARGENTINA

Universidad Nacional de R’o Cuarto

CASTELAR, ARGENTINA

Instituto Nacional de Tecnolog’a Agropecuaria

 

Path coefficient analysis of ÒMal de R’o CuartoÓ disease components.

--Dallo, MD; Iba–ez, MA; Borghi, ML; Girardi, VN; Bonamico, NC; Salerno, JC; D’az, DG; Di Renzo, MA.

 

INTRODUCTION

The ÒMal de R’o CuartoÓ (MRC) is considered the most important viral disease of corn in Argentina, causing increase damage and high yield losses to susceptible cultivars (Lenard—n, SL et al., Plant Dis 82:448, 1998; Di Renzo, MA, et al., J Agr Sci 139:47-53, 2002). The disease is caused by MRC virus which belongs to the genus Fijivirus, within the family Reoviridae and is transmitted by the planthopper Delphacodes kuscheli Fennah (Hemiptera: Delphacidae) (Nome, SF et al., Phytopathol Z 101:7-15, 1981; Ornaghi, JA et al., J Gen Breed 47:227-282, 1993). Its symptomatology depends of the plant phenological state when it is infected, the genotype and the environmental conditions where the culture grown. This includes stunting, short internodes, galls or enations on the abaxial side of the leaves, size reduction and malformation of spikes, ears and leaves (Lenard—n, SL et al, SL Lenard—n Ed., IFFIVE, INTA, JICA, 1999; Di Renzo, MA et al., J Agric Sci 142:289-295, 2004). Since both economic and environmental reasons prevent a raise in agrochemicals usage, future cereal improvement will rely on germplasm that optimize present genetic tolerance to plant pathogens (Abeledo, LG et al., Euphytica 130:325-334, 2003). This is an effective way of both increasing and stabilizing production in affected areas.

Selection for one trait usually affects several traits. The correlated response to selection is a change in one or more traits due to selection of another as a result of genetic relations between them. Genetic correlations are useful to decide on selection strategies since they express the relative importance of pleiotropy and linkage between loci (Kang, MS et al., Crop Sci 23:643-647, 1983; Kang, MS, Applied quantitative genetics. MS Kang Publisher, Baton Rouge, LA, 1994). The genetic correlations can be estimated from phenotypic values, which are influenced by type of gene action, environmental effects and genotype x environment interactions (Falconer and Mackay, Introduction to quantitative genetics. Longman Technical, Essex, UK, 1996).

Nevertheless, while correlation coefficient only measures the association magnitude between variables, the path coefficient analysis (Wright, S, J Agric Res 20:557-587, 1921; Wright, S, Ann Math Stat 5:161-215, 1934) allows dissecting the correlation between them into effects, direct and indirect. This method has been commonly used in crop breeding studies to establish the relationships between grain yield and its contributing components (Mohammadi, SA et al., Crop Sci 43:1690-1697, 2003). In addition, there are many references of its application in plant pathology studies (Van Bruggen and Arneson, Phytopathology 76:874-878, 1986; Nayak, P et al., J Phytopathol 119:312-318, 1987; Neher, DA et al., Plant Dis 77:1106-1111, 1993; Birhman and Singh, Ann Appl Biol 127:353-362, 1995; Desprez-Loustau and Wagner, Eur J Plant Pathol 103:653-665, 1997; Garcia, D et al., Eur J Forest Pathol 29:323-338, 1999). Therefore, it would be useful to get information regarding direct and indirect relationships among MRC disease tolerance and different related traits. The objectives of this work were: i) to determine phenotypic and genetic correlation coefficients among MRC disease tolerance and different related traits and ii) to present a path coefficient analysis to show how these traits affect MRC disease tolerance.

 

MATERIALS AND METHODS

Plant material

One hundred and forty four F2:6 recombinant inbred lines (RIL) were developed from a cross between Mo17, a susceptible dent line, and BLS14, a partially resistance red flint line. The RILs were assessed to MRC disease and parental genotypes were used as controls in each plot.

Field trials

The RILs were tested during the 2004 summer cycle in field experiments at two locations belonging to the endemic area: R’o Cuarto (R4) and Sampacho (S4). The field trials were conducted in a randomized complete block design with two replicates of single-row plots. Each plot consisted of 3.0 m rows with 0.7 m spacing. Plants were thinned to a distance of 0.15 m in the row. The sowing date determination was made through insect vector monitoring during spring, in order to Delphacodes kuscheli population reach the peak during early stages of maize development.

Symptoms observed and scored

Disease symptoms were scored 60-70 days after male flowering. Individual plants from each plot were phenotypically screened for traits related to MRC disease: plant height (PH), internodes (IN), enations (EN), flag leaf edge (LE), width (LW) and length (LL), panicle (PA) and ear (EA). A disease severity grade was estimated for each plant according to a 0 - 3 scale proposed by Ornaghi et al. (Maydica 44:219-223, 1999). The response variable is the disease severity index (DSI) based on the disease severity grades of individual plants. DSI was calculated for each plot and used to rate RIL for their reaction to MRC disease according to Grau et al. (Plant Dis 66:506–508, 1982). The details of the rating for MRC severity were described in Di Renzo et al. (J Agr Sci 139:47-53, 2002).

Statistical analysis

Data from each trial location were subjected to variance and covariance analyses using the PROC GLM procedure of SAS (SAS Institute ver 9.1.3). Genetic and phenotypic correlation coefficients among DSI and traits related to MRC disease were determined from variance and covariance components. The path coefficient analysis was performed to calculate direct and indirect effects among the response variable, DSI, and MRC related predictor variables. Direct and indirect path coefficients were calculated as firstly proposed Wright (J Agric Res 20:557-587, 1921; Ann Math Stat 5:161-215, 1934), and then Dewey and Lu (Agron J 51:515-518, 1959) and Li (Path analysis: a primer. Boxwood Press, Pacific Grove, CA, 1975). For each trait, direct path coefficient was solved by means of PROC IML (SAS Institute ver 9.1.3). The product of appropriate correlation coefficient (r) and path coefficient values provides the indirect path coefficient. Correlations and path coefficient analyses based on genetic values define more precisely what factors affect DSI genetically. Previous to analysis, the original data was logarithmically transformed to satisfy the assumption of additivity among the components. The presence of multicollinearity among variables was measured using the variance inflation factor (VIF) and the condition number (CN). Residual effects and determination coefficients were estimated according to Kang (Applied quantitative genetics. MS Kang Publisher, Baton Rouge, LA, 1994). The criterion followed to evaluate the extent of effects of MRC related traits on DSI magnitude was according to Board et al. (Crop Sci 37:879-884, 1997).

 

RESULTS AND DISCUSSIONS

Phenotypic and genetic correlations coefficients among all pairs of traits are shown in Table 1. In general, phenotypic and genetic correlation coefficients agreed in sign and the magnitude of the phenotypic correlation coefficient was practically the same that the genetic correlation coefficient indicating that the influence of environment on these relationships was little or negligible. However, in most cases, the genetic correlation estimates between DSI and MRC disease related traits were slightly high showing that they are genetically associated or that they are physiologically related (Sidwell, RJ et al., Crop Sci 16:650-654, 1976). At R4 location, EN had the highest positive phenotypic correlation on DSI (0.94) followed by IN (0.91) and EA (0.64) but, IN had the highest positive genetic correlation on DSI (0.96) followed by EN (0.95) and EA (0.69). However at S4 environment, both phenotypic and genetic correlation coefficient of EN on DSI showed the highest positive values (0.96 and 0.98, respectively) followed by IN (0.90 and 0.94, respectively) and PA (0.69 and 0.77, respectively).

 

Table 1. Phenotypic and genetic correlations among all pairs of traits at R4 and S4 environments.

 

PA

LL

LW

LE

EN

IN

EA

DSI

PA

 

 

0.490

0.535

 

0.408

0.399

 

0.237

0.225

 

0.576

0.596

 ++

0.552

0.581

 

0.255

0.298

 

0.591

0.611

 ++

 

 

 

 

 

 

 

 

 

LL

0.577

0.703

 

 

 

0.411

0.412

 

0.175

0.128

 

0.353

0.358

 

0.309

0.363

 ++

0.019

0.002

 

0.352

0.375

 ++

 

 

 

 

 

 

 

 

 

LW

0.444

0.320

 

0.362

0.340

 

 

 

0.119

0.042

 

0.155

0.092

 

0.148

0.137

 

-0.044

-0.143

 

0.189

0.166

 

 

 

 

 

 

 

 

 ++

 

LE

0.324

0.350

 

0.480

0.530

 ++

0.128

-0.038

 

 

 

0.534

0.558

 

0.518

0.582

 ++

0.596

0.679

 ++

0.541

0.556

 

 

 

 

 

 

 

 

 

 

EN

0.666

0.762

 

0.622

0.747

 ++

0.226

0.243

 

0.566

0.714

 

 

 

0.867

0.913

 ++

0.600

0.648

 ++

0.937

0.955

 ++

 

 

 

 

 

 

 ++

 

 

IN

0.616

0.697

 ++

0.530

0.646

 

0.212

0.221

 

0.483

0.603

 

0.849

0.891

 ++

 

 

0.559

0.651

 ++

0.914

0.961

 ++

 

 

 

 

 ++

 

 

 

 

EA

0.427

0.469

 

0.233

0.272

 

0.049

-0.016

 

0.282

0.331

 ++

0.528

0.600

 

0.486

0.570

 

 

 

0.642

0.691

 ++

 

 

 

 

 

 

 

 

 

DSI

0.685

0.774

 ++

0.597

0.707

 ++

0.232

0.240

 

0.532

0.660

 

0.961

0.979

 ++

0.898

0.941

 ++

0.549

0.628

 ++

 

 

 

 

 

 

 

 

 

 

 

 

  R4 = R’o Cuarto 2004; S4 = Sampacho. DSI = disease severity index; PA = panicle; LL = leaf lenght; LW = leaf width; LE = leaf edge; EN = e-nations; IN = internodes; EA = ear.

++, + Significant at the 0.01 and 0.05 levels, respectively, by bootstrap method.

¤ Values above/below the diagonal: correlations among all pairs of traits at R4 and S4 environments, respectively. Genetic values are shown in bold.

 

Fig. 1. Path diagram showing causal relationships between disease severity index (DSI) and Mal de R’o Cuarto (MRC) disease related traits. One-headed arrow represents direct paths and double-headed arrows represent correlations (r).

 

 

 

Because of the similarity between phenotypic and genetic correlations, the phenotypic and genetic path coefficients were also quite similar. The path diagram based both phenotypic and genetic correlation coefficients were carried out according to shown in Fig. 1. The almost equal magnitude of EN and IN direct effects on DSI at both phenotypic and genetic levels indicated that these effects were under genetic control with unimportant environmental effects. However, at R4 and S4 environments, the genetic direct effects of EN and IN increased hardly over their respective phenotypic direct effects which suggest a negative relationship between DSI and environmental effects. In general, EN and IN were the primary and secondary direct determinants of DSI over both environments at the phenotypic and genetic levels (Table 2). Nevertheless, at R4 location, IN showed the highest positive genetic direct effect on DSI (0.490) followed by EN (0.445). Similarly, the large positive phenotypic and genetic direct effects of EN and IN on DSI was counterbalanced by a small phenotypic and genetic indirect effects via the remaining traits included in the analysis. In most cases, the remaining traits had negligible phenotypic and genetic direct effects but the phenotypic and genetic indirect effects were intermediate via EN and IN on both environments (Table 2).


 


Table 2. Phenotypic and genetic path coefficient analysis of DSI and MRC components at R4 and S4 environments.

Pathway

R4

S4

DSI vs. PA

 

 

 

 

Direct effect

0.032

-0.010

0.052

0.053

Indirect effect via

 

 

 

 

      LL

0.012

0.011

-0.004

-0.032

      LW

0.016

0.032

-0.001

0.000

      LE

-0.005

-0.020

-0.004

-0.011

      EN

0.290

0.266

0.454

0.532

      IN

0.213

0.285

0.173

0.219

      EA

0.033

0.047

0.016

0.013

      Correlation

0.591

0.611

0.685

0.774

DSI vs. LL

 

 

 

 

Direct effect

0.025

0.021

-0.007

-0.046

Indirect effect via

 

 

 

 

      PA

0.016

-0.005

0.030

0.038

      LW

0.016

0.033

-0.001

0.000

      LE

-0.004

-0.011

-0.007

-0.017

      EN

0.178

0.159

0.424

0.522

      IN

0.119

0.178

0.149

0.203

      EA

0.003

0.000

0.008

0.007

      Correlation

0.352

0.375

0.597

0.707

DSI vs. LW

 

 

 

 

Direct effect

0.038

0.080

-0.002

-0.001

Indirect effect via

 

 

 

 

      PA

0.013

-0.004

0.023

0.017

      LL

0.010

0.009

-0.002

-0.016

      LE

-0.003

-0.004

-0.002

0.001

      EN

0.078

0.041

0.154

0.169

      IN

0.057

0.067

0.060

0.070

      EA

-0.006

-0.023

0.002

0.000

      Correlation

0.189

0.166

0.232

0.240

DSI vs. LE

 

 

 

 

Direct effect

-0.022

-0.089

-0.014

-0.031

Indirect effect via

 

 

 

 

      PA

0.008

-0.002

0.017

0.019

      LL

0.004

0.003

-0.003

-0.024

      LW

0.005

0.003

0.000

0.000

      EN

0.269

0.249

0.386

0.498

      IN

0.200

0.285

0.136

0.190

      EA

0.077

0.108

0.010

0.009

      Correlation

0.541

0.556

0.532

0.660

DSI vs. EN

 

 

 

 

Direct effect

0.504

0.445

0.681

0.698

Indirect effect via

 

 

 

 

      PA

0.019

-0.006

0.035

0.041

      LL

0.009

0.008

-0.004

-0.034

      LW

0.006

0.007

-0.001

0.000

      LE

-0.012

-0.049

-0.008

-0.022

      IN

0.334

0.447

0.238

0.280

      EA

0.078

0.103

0.019

0.016

      Correlation

0.937

0.955

0.961

0.979

DSI vs. IN

 

 

 

 

Direct effect

0.385

0.490

0.281

0.315

Indirect effect via

 

 

 

 

      PA

0.018

-0.006

0.032

0.037

      LL

0.008

0.008

-0.003

-0.030

      LW

0.006

0.011

0.000

0.000

      LE

-0.011

-0.052

-0.007

-0.019

      EN

0.436

0.407

0.578

0.622

      EA

0.073

0.103

0.018

0.016

      Correlation

0.914

0.961

0.898

0.941

DSI vs. EA

 

 

 

 

Direct effect

0.130

0.158

0.036

0.027

Indirect effect via

 

 

 

 

      PA

0.008

-0.003

0.022

0.025

      LL

0.000

0.000

-0.002

-0.012

      LW

-0.002

-0.011

0.000

0.000

      LE

-0.013

-0.060

-0.004

-0.010

      EN

0.302

0.288

0.359

0.419

      IN

0.215

0.319

0.137

0.179

      Correlation

0.642

0.691

0.549

0.628

Coefficient of determination

0.931

0.971

0.951

0.985

Residual effect

0.262

0.170

0.222

0.123

 

  R4 = R’o Cuarto 2004; S4 = Sampacho. DSI = disease severity index; PA = panicle; LL = leaf lenght; LW = leaf width; LE = leaf edge; EN = enations; IN = internodes; EA = ear.

¤ Genetic values are shown in bold.

The phenotypic and genetic path coefficients for DSI accounted for a large proportion of phenotypic and genetic variation on both environments as indicated by a large coefficient of determination and by the corresponding small residual effect. In the phenotypic and genetic path analysis, the residual effect represents the failure of the estimated genetic correlations among the variables to account for the total genetic variation in a trait (Sidwell, RJ et al., Crop Sci 16:650-654, 1976).

The most important components of DSI, if we just consider phenotypic and genetic correlation coefficients, were EN, IN, EA, PA and LE. Nevertheless, phenotypic and genetic path coefficient analysis showed that among these traits only EN and IN were the most important to increase DSI. The direct effects of EA, PA and LE were little or negligible. The results of this study suggest that EN and IN are the best traits in determining DSI on both environments and may be useful as an indirect selection criteria in breeding and selection programs related to MRC disease.

 

REFERENCES

Abeledo, L.G., D.F. Calderini and G.A. Slafer. 2003. Genetic improvement of barley yield potential and its physiological determinants in Argentina (1944-1998). Euphytica 130: 325-334.

Birhman, R.K. and B.P. Singh. 1995. Path-coefficient analyses and genetic parameters of the components of field resistance of potatoes to late blight. Ann Appl Biol. 127: 353-362.

Board, J.E., M.S. Kang and B.G. Harville. 1997. Path analyses identify indirect selection criteria for yield of late-planted soybean. Crop Sci. 37: 879-884.

Desprez-Loustau, M.L. and K. Wagner. 1997. Components of maritime pine susceptibility to twisting rust – A path coefficient analysis. Eur J Plant Pathol. 103: 653-665.

Dewey, D.R. and K.H. Lu. 1959. Correlation and path coefficient analysis of components of crested wheat grass seed production. Agron J. 51: 515-518.

Di Renzo, M.A., N.C. Bonamico, D.D. D’az, J.C. Salerno, M.A. Iba–ez and J.J. Gesumaria. 2002. Inheritance of resistance to Mal de R’o Cuarto (MRC) disease in Zea mays (L.). J Agr Sci. 139: 47-53.

Di Renzo, M.A., N.C. Bonamico, D.D. D’az, M.A. Iba–ez, M.E. Faricelli, M.G. Balzarini and J.C. Salerno. 2004. Microsatellite markers linked to QTL for resistance to Mal de R’o Cuarto disease in Zea mays (L.). J Agric Sci. 142: 289-295.

Falconer, D.S. and T.F.C. Mackay. 1996. Introduction to quantitative genetics, 4th ed. Longman Technical, Essex, UK.

Garcia, D., V. Troispoux, N. Grange, F. Rivano and J. D'Auzac. 1999. Evaluation of the resistance of 36 Hevea clones to Microcyclus ulei and relation to their capacity to accumulate scopoletin and lignins. Eur J Forest Pathol. 29: 323-338.

Grau, C.R., V.L. Radke and F.L. Gillespie. 1982. Resistance of soybean cultivars to Sclerotinia sclerotiorum. Plant Dis. 66: 506–508.

Kang, M.S., J.D. Miller and P.Y.P. Tai. 1983. Genetic and phenotypic path analyses and heritability in sugarcane. Crop Sci. 23: 643-647.

Kang, M.S. 1994. Applied quantitative genetics. M.S. Kang Publisher, Baton Rouge, LA.

Lenard—n, S.L., G.J. March, S.F. Nome and J.A. Ornaghi. 1998. Recent outbreak of ÒMal de R’o Cuarto VirusÓ on corn in Argentina. Plant Dis 82: 448.

Lenard—n, S.L., G.J. March y J.A. Ornaghi, 1999. Virus del Mal de R’o Cuarto en ma’z. En: Lenard—n, S.L. (Ed.). Enfermedades causadas por virus y fitoplasmas en cultivos extensivos intensivos. Proyecto de Investigaciones en Fitovirolog’a. IFFIVE, INTA, JICA. Hoja Informativa. Ma’z N¼ 2, 10 pp.

Li, C.C. 1975. Path analysis: a primer, 3rd ed. Boxwood Press, Pacific Grove, CA.

Mohammadi, S.A., B.M. Prasanna and N.N. Singh. 2003. Sequential path model for determining interrelationships among grain yield and related characters in maize. Crop Sci. 43: 1690-1697.

Nayak, P., A.V. Suriya Rao and N.K. Chakrabarti. 1987. Components of resistance to bacterial blight disease of rice. J Phytopathol. 119: 312-318.

Neher, D.A., C.D. McKeen and J.M. Duniway. 1993. Relationships among phytophthora root rot development, P. parasitica populations in soil, and yield of tomatoes under commercial field conditions. Plant Dis. 77: 1106-1111.

Nome, S.F., S.L. Lenard—n, B.C. Raju, I.G. Laguna, S.K. Lowe and D. Docampo. 1981. Association of reovirus-like particles with ÒEnfermedad de R’o IVÓ of maize in Argentina. Phytopathol Z. 101: 7-15.

Ornaghi, J.A., G.T. Boito, G. Sanchez, G.J. March and J.E. Beviacqua. 1993. Studies on the population of Delphacodes kuscheli Fennah in different years and agricultural areas. J Gen Breed. 47: 227-282.

Ornaghi, J.A., G.J. March, G.T. Boito, A. Marinelli, J.E. Beviacqua, J. Giuggia and S.L. Lenard—n. 1999. Infectivity in natural populations of Delphacodes kuscheli vector of ÒMal de R’o CuartoÓ virus. Maydica 44: 219-223.

SAS Institute Inc. 2005. SAS/STAT user«s guide, Version 9.1.3. SAS Institute Inc., Cary, NC.

Sidwell, R.J., E.L. Smith and R.W. McNew. 1976. Inheritance and interrelationships of grain yield and selected yield-related traits in a hard red winter wheat cross. Crop Sci. 16: 650-654.

Van Bruggen, A.H.C. and P.A. Arneson. 1986. Path coefficient analysis of effects of Rhizoctonia solani on growth and development of dry beans. Phytopathology 76: 874-878.

Wright, S. 1921. Correlation and causation. J Agric Res. 20: 557-587.

Wright, S. 1934. The method of path coefficients. Ann Math Stat. 5: 161-215.

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