easygenetics

Candice Sammons
Piyush Chaudhari

2018-04-24

Contents

  • Introduction to easygenetics
  • Installation
    • GitHub
    • Posit Connect Cloud
  • User Interface
  • List of Functions
  • Using functions in easygenetics
  • Significance of easygenetics
  • Additional Details
  • Session Info
  • Contact Details

Introduction to easygenetics

  • easygenetics is a front end genetic data analysis platform in R programming environment
  • It is a R package, consisting of functions which can be accessed and played with via a graphical interface
  • The aim of this package is to create an easily accessible platform for non-statisticians and non-programmers to work on with minimal coding
  • The user will have the minimum task of just loading the package and running the required function
  • The input variables can be entered in the GUI, and results are available almost instantaneously, which improves the accessibility for non-programmers

Installation

  1. GitHub

The r package devtools is needed to install the easygenetics package from GitHub

devtools can be installed using

install.packages("devtools", repos = "https://cloud.r-project.org/")

The downloaded binary packages are in
    /var/folders/1j/qb3qtyq94gxfdcs5wx9wt4zr0000gn/T//RtmpRnOu7G/downloaded_packages

Then the easygenetics package can be installed using install_github() function from devtools package as below:

devtools::install_github("pc-piyush/easygenetics")

Installation

  1. Posit Connect Cloud

The Shiny app can be accessed using the link - easygenetics

The app is made using shiny package in R and hosted on Posit Connect Cloud

User Interface

Fig 1: easygenetics User Interface

  • The Function Bar consists of the index of functions which makes it easier to access when you are using the package

  • The Action Area allows you to manipulate the parameters and observe and visualize the results parallely

  • The Results area allows you to visualize the results

  • The Run button lets you submit your values and observe the results

List of Functions

  • Simulating Genotypes at a SNP over Multiple Generations : Simulate_Genotypes()
  • Genotypic Values, Breeding Values, and Dominance Deviation Table : Breeding_Values()
  • Interval Mapping : Interval_Mapping()
  • Multiple QTL Mapping : Multiple_QTL_Mapping()
  • Simulating Genotype of Parent by Child- Estimating Covariance : Covariance_Estimation()
  • Haseman-Elston Regression : Haseman-Elston_Regression()

Simulate_Genotypes()

The function located on the first tab titled Simulate_Genotypes() is used to simulate genotypes at a SNP over multiple generations with the assumption that the mutation only happens one way

Input

  • The arguments that are required in order for this function to properly run is the initial sample size , mutation rate , average number of offspring per generation and number of generations
  • The Enter Seed is a numeric value that allows the results to be reproducible
  • All of the required arguments needed in order for the function to work are numeric values
  • The initial sample size is the starting population size at the beginning of the simulation
  • The mutation rate is the frequency of the target mutation in one gene over time
  • The average number of offspring per generation is the mean growth in each generation

Output

  • The output of this function produces two plots
    • The first plot shows the growth in population size over generation
    • The second plot provides the allele frequency over generation

These plots are both determined by the user’s arguments that are defined above.

Simulate_Genotypes()

Fig 2: Simulate_Genotypes()

Breeding_Values()

The second tab is titled Breeding_Values(), which calculates values based on the inputted arguments

Input

  • The inputs for this function are all numeric
  • A1A1 is the number of observations of the phenotypic means of the specified homozygous genotype
  • A1A2 is the number of observations of the phenotypic means of the heterozygous genotype, and
  • A2A2 is the number of observations of the phenotypic means of the specified homozygous genotype
  • The Frequency of Allele A1 is the rate at which allele A1 occurs, and the Frequency of Allele A2 is the rate at which allele A2 occurs

Output

  • The output of this function provides

    • Table that includes midpoint (m) , population mean (M) , additive effect (a) , dominance effect (d)
    • Another table with the frequency , genotypic value , breeding value and dominance deviation of each genotype

Breeding_Values()

Fig 3: Breeding_Values()

Breeding_Values()

Glossary

  • Midpoint is the middle value of the phenotypic means between the two homozygous genotypes

  • Expected mean phenotype is calculated using the additive effect and dominance effect , where

    • additive effect is the distance between the midpoint and the phenotypic means of the homozygous genotypes, and
    • dominance effect is the distance between the midpoint and phenotypic mean of the heterozygous genotype

  • Population mean is calculated from the additive effect , dominance effect and the frequencies of each genotype

  • Frequency is the rate at which each genotype occurs

  • Genotypic value is the additive effect and dominance effect estimates

  • Breeding value is the sum of average effects of the alleles

  • Dominance Deviation is the difference between genotypic value and breeding value

Interval_Mapping()

Interval_Mapping() is the function located on the third tab of the app - Interval mapping is currently the most common method used for QTL mapping - Its used by assuming the presence of a single QTL, and estimates the QTL position between two chosen markers after considering each locus one at a time

Input and Output

  • The Enter Seed is a numeric value that allows the results to be reproducible

  • Function is a select list input with the options

    • Properties of Cross
    • Plot
    • LOD Thresholds, and
    • LOD Intervals

    The default in this list is Properties of Cross and the output changes depending on this selection

    • If Properties of Cross is selected and all the other parameters are entered and selected, the output displays the summary of a data simulation for a QTL experiment. This summary includes information such as the type of cross , number of individuals , number of phenotypes , total markers , percent genotyped and so on

    • If Plot is selected and all the other parameters are entered and selected, the output displays an interval mapping plot. Depending on what is selected under Interval Mapping Method determines what plot is displayed

    • If the option LOD Thresholds is selected, the output provides the summary of a genome scan with a single QTL model. This includes the interval mapping threshold that shows which chromosome , position and LOD value where the LOD threshold is located

    • Selecting LOD Interval under Function produces an output that shows the LOD support interval for the chromosome

Interval_Mapping()

Input and Output

  • The output described above is determined based on the other parameters that the user provides

  • This includes the length of chromosome , the number of markers , the number of individuals , the number of permutations , the position of QTL1 and QTL2 and the effect of QTL1 and QTL2 , which are all numeric values

  • Equal Spacing , Type of Cross , No. of QTLs and Interval Mapping Method are all select list inputs due to the restrictions of the function

    • Equal Spacing provides the option of TRUE or FALSE that determines the presence of equal spacing in the genetic map simulation
    • Type of Cross specifies what kind of population to simulate, and has the option of bc which stands for backcross and f2 for intercross
    • The choices for No. of QTLs are 1 and 2 due to function constraints
    • Interval Mapping method is used to determine which plot is shown. The options for the different methods are EM , HK , EHK and All

Interval_Mapping()

Fig 4: Interval_Mapping()

Multiple_QTL_Mapping()

Multiple_QTL_Mapping() is the next function in the tab and is used for multiple QTL mapping by the Haley-Knott Regression

  • The downside of mapping one QTL at a time is that a second QTL might be missed, or ghosted
  • By performing multiple QTL mapping, more QTLs can be identified on the same chromosome

Input

To execute this function, there are parameters that need to be set, which will be described below

  • The Enter Seed is a numeric value that allows the results to be reproducible
  • The Type of Function is a select list input with the options of
    • Genetic Map
    • Single QTL
    • Two QTLs
    • Heat Map-Two QTLs Comparison
    • Two QTLs Comparison
    • Composite Interval Mapping
  • The default is set to Genetic Map

Multiple_QTL_Mapping()

Output

Depending on what is selected and what the other inputs are, determines the output displayed

  • By selecting Genetic Map , the output shows a plot marker map
  • The Single QTL option plots the LOD score for each map position in centimorgans of a single QTL, which is the QTL the user clarifies in QTL1. A summary of this using Haley-Knott regression is presented below the plot
  • The same output is shown for when Two QTLs is selected, except now the method uses a two-QTL model
  • Choosing Heat Map-Two QTLs Comparison changes the output into a heat map of the QTLs
  • With Composite Interval Mapping a plot is given with the black curve representing single interval mapping where the QTLs are not clearly distinguished and the blue curve is the composite interval curve. Below the plot the LOD threshold of the composite interval curve is provided in summary.

The rest of the parameters are similar to the ones in the Interval_Mapping function. In order for the output to be present, there needs to be a numeric value entered into the length of chromosome , the number of markers , the number of individuals , the position of QTL1 and QTL2 , and the effect of QTL1 and QTL2. Equal Spacing has a choice between TRUE and FALSE that determines the presence of equal spacing in the genetic map simulation. Type of Cross has options bc for backcross and f2 for intercross to determine what kind of population to simulate.

Multiple_QTL_Mapping()

Fig 5: Multiple_QTL_Mapping()

Covariance_Estimation()

  • The Covariance_Estimation function is used to estimate variance between a parent and offspring

Input

  • The user must specify the number of generations \(n\) as a numeric value under parameter Enter n

    The covariance is estimated through simulation, using a normal distribution to randomly generate the two parental allelic effects and a binomial distribution to randomly determine the transmitted allele. The transmitted allele assists in determining the offspring’s phenotype. Before estimating the covariance the appropriate errors are added to the parent and offspring, such as the environmental error.

  • The Enter Seed is a numeric value that allows the results to be reproducible.

Output

The output displayed is the covariance estimation between parents and offsprings established from \(n\)

Covariance_Estimation()

Fig 6: Covariance_Estimation()

Haseman-Elston_Regression()

The Haseman-Elston Regression function is used to obtain a series of plots which represent the relationship between the phenotypes of two siblings.For each sibling pair, the proportion of alleles IBD are determined at a given locus. The regression is done on squared difference between the sibling trait values on the proportion of alleles shared IBD. The variance component comes from genetic and environmental factors.

If \(y\)\(i\) denote the outcome variable for subject \(i\), then the Haseman-Elston regression model is-

\(y\)\(i\) = \(\mu\) + \(\alpha\)\(i\) + \(\alpha\)*\(i\) + \(\epsilon\)\(i\)

where \(\alpha\)\(i\) and \(\alpha\)* represents the additive genetic variance contribution from QTLs from inside and outside the interval. \(\alpha\), \(\alpha\)* and \(\epsilon\) are uncorrelated random errors, and the phenotypic variance is the sum of variance of these 3 random terms. The sum of \(\sigma_\alpha^2\) and \(\sigma_\alpha^{*2}\) is the total genotypic variance, whereas \(\sigma_\epsilon^2\) represents the error variance, which is the difference between the total variance and sum of genotypic and environmental variance.

Haseman-Elston_Regression()

Input

  • The Enter Seed is a numeric value that allows the results to be reproducible

  • Type of Plot is the select list input where we can select the type of plot we want in the result window. The options are Histogram of IBD , IBD vs Squared differences between phenotypes , Phenotype plot- Sibling 1 vs Sibling 2 and Phenotype plot- Sibling 1 vs Sibling 2 for IBD= 0,1,2. An additional option is All , which shows all the above mentioned plots together

  • Number of Families is a numeric input variable which indicates the number of families on which the regression is to be performed

  • Total Variance , Genetic Variance and Environmental Variance represent the Total Phenotypic Variance and the contribution of Genetic and Environmental component respectively

  • Minor Allele Frequency is the frequency of the minor allele in the population.

Output

The default output plot type is All , which produces all plots- Histogram of IBD , IBD vs Squared differences between phenotypes , Phenotype plot- Sibling 1 vs Sibling 2 and Phenotype plot- Sibling 1 vs Sibling 2 for IBD= 0,1,2. Histogram of IBD shows the frequency distribution of population which belongs to each IBD category. IBD vs Squared differences between phenotypes plots the squared difference between phenotypic values of siblings by the three IBD values- 0,1 and 2. Phenotype plot- Sibling 1 vs Sibling 2 shows the overall scatter plot of phenotypic values of sibling 1 and sibling 2. Phenotype plot- Sibling 1 vs Sibling 2 for IBD= 0,1,2 represent the subset of the overall phenotypic plot segmented by IBD values 0,1 and 2.

Haseman-Elston_Regression()

Fig 7: Haseman-Elston_Regression()