Varibility analysis in R - III: Genotypic path analysis

    The blog is third and last part of a series of three blogs related to variability analysis using R software. The blog demonstrates use of variability package available in R for carrying out genotpyic path analysis among traits in one line of code. (Reading time 9 min.)

    Before heading further to read this blog it is highly recommended to go through first blog Variability analysis in R - I: Genetic parameters .Reading the first blog will give an idea about how to arrange the data in excel sheet, import the data in RStudio and install the package named variability required for the analysis.

    For path analysis we need have one dependent variable (trait) and other as independent variables. We generally choose yield or yield attributing trait as a dependent variable. In our case we will use GRY as a dependent variable. Make sure the dependent variable is kept in last column of data set.

R Script for Genotypic correlation

Lines in blue are script. Type these blue lines and press Ctrl+Enter run the lines.

Open a new script by clicking on New and than selecting R script or by pressing Ctrl+Shift+N.

library(variability)

library() will load the variability package

?geno.path

In variability package we have function named geno.path() to carryout genotypic path analysis. To know more about the function run the below mentioned line.

Run the below line to carry out genotypic path analysis

geno.path(vardata1[8],vardata1[3:7],vardata1$Genotypes,vardata1$Replication)

geno.path is the function for genotypic path analysis

vardata1[8] here square bracket are used for indexing. First we need to place dependent variable. GRY is our dependent variable and located on eighth column, so 8 is written in square bracket.

vardata1[3:7] here square bracket are used for indexing. In second position we need to place independent variables which are located between third to seventh column, so 3:7 is written in square bracket.

vardata1$Genotypes is used to use genotypes as an input. R will look for a column named Genotypes in vardata1 dataset

vardata1$Replication is used to use replication as an input. R will look for a column named Replication in vardata1 dataset

Output

$Effects
           PH       LL      MRP      GRW     DRWR
PH    0.01767 -0.02402 -0.00270  0.42371  0.00988
LL    0.01385 -0.03063  0.00193  0.15121  0.00716
MRP  -0.00394 -0.00487  0.01211  0.76361  0.00687
GRW   0.00769 -0.00476  0.00950  0.97397  0.01356
DRWR -0.01053  0.01323 -0.00501 -0.79621 -0.01659

$Residual
      [,1]
[1,] 1e-04

The direct effects are on diagonal and indirect effects are on off-diagonal of the output matrix

How to check if the path analysis output is right or not?

Path analysis is basically breaking down the correlation between dependent variable and one independent variable into direct effect  and indirect effects.

For example,

Genotypic correlation (PH, GRY) = Direct effect of PH + Sum of indirect effects of PH via LL, MRP, GRW and DRWR

Lets check it out

Genotypic correlation (PH, GRY) = 0.4245 (obtained from second blog of variability series)

1. Direct effect of PH = 0.01767
2. Indirect effect of PH via LL = -0.02402
3. Indirect effect of PH via MRP = -0.00270
4. Indirect effect of PH via GRW = 0.42371
   
5. Indirect effect of PH via DRWR = 0.00988

 Summation of these five values gives 0.42454 which indicates our calculation is right.

Hope you find this blog helpful

If you are using the variability R package in research work you can cite it as:
Raj Popat, Rumit Patel and Dinesh Parmar (2020). variability: Genetic Variability Analysis for Plant Breeding Research. R package version 0.1.0.https://CRAN.R-project.org/package=variability

 

Topics you might be interested:

Variability analysis in R - I: Genetic parameters

Varibility analysis in R - II: Genotypic & Phenotypic correlation