vignettes/RMeDPower_tutorial.md

In gladstone-institutes/CalcPower: Power Calculation for biological experiments

RMeDPower_tutorial

body{ font-size: 16pt; }

Introduction

This function uses simulation to perform power analysis. It is designed to explore the power of biological experiments and to suggest an optimal number of experimental variables with reasonable power. The backbone of the function is based on simr package, which fits a fixed effect or mixed effect model based on the observed data and simulates response variables. Users can test the power of different combinations of experimental variables and parameters.

Installation

library(devtools)
install_github('gladstone-institutes/RMeDPower', build_vignettes=TRUE)

First few lines of the input data

head(RMeDPower_data1)
#>   experiment plate  line classification feature1 feature2   feature3 feature4
#> 1       exp7     1 line1              0       62 3261.238   16.69571 118.4444
#> 2       exp7     1 line1              0       77 2033.726   54.10482 114.0946
#> 3       exp7     1 line1              0       56 1935.731 -100.00000 122.7329
#> 4       exp7     1 line1              0       54 1533.812 -100.00000 118.7066
#> 5       exp7     1 line1              0       19 1437.420 -100.00000 117.0279
#> 6       exp7     1 line1              0       30 1899.334 -100.00000 117.9523
#>   feature5
#> 1       10
#> 2       14
#> 3       13
#> 4       13
#> 5       19
#> 6       11

Ex1. Varaince estimation and power calculation from pilot data

Levels will be increased to max(observed level)x5. For example, there are 9 experiments in the original data:

exp1, exp2, exp3, …, exp9

and we will test 9 x 5 experiments

code example:

calculate_power(data=RMeDPower_data1,power_curve=1,
                 variance_estimate_from="data",condition_variable="classification",
                 experimental_variable=c("experiment","plate","line"), response_variable="feature1",
                 nsimn=10, target_parameters="experiment", levels=1)  

* power_curve=1 : to get a power curve that calculates power for different levels of the target parameter

* variance_estimate_from =“data” : to estimate variance values from the input data

* condition_variable : cell status variable (ex control/case)

* experimental_variable : variables related to the experimental design

* response_varaible: phenotype

* nsimn=10 : 10 iterations for power calculation ( try at least 100 or 1000 to get high accuracy)

* target_parameters=“experiment” : to explore “experiment”

* levels = 1: to explore different levels of the target parameter

Ex2. Varaince estimation and power calculation from pilot data

Sample sizes will be increased to max(observed sample size)x5. For example, when there are max 71 samples per cell line in the original data:

table(RMeDPower_data1$line)
#> 
#> line1 line2 line3 line4 line5 line6 line7 line8 
#>    42    40    37    51    71    37    42    20

and we will test max 71 x 5 samples per cell line

code example:

calculate_power(data=RMeDPower_data1,power_curve=1,
                 variance_estimate_from="data",condition_variable="classification",
                 experimental_variable=c("experiment","plate","line"), response_variable="feature1",
                 nsimn=10, target_parameters="line", levels=0)  

* levels = 0 : to explore different sample sizes of the target parameter

Ex3. User determined level count and output file name

Varaince estimation and power calculation (for a single level size) from pilot data with user determined level count and output file name. We will do Ex1 with max level size 15

code example:`

calculate_power(data=RMeDPower_data1,power_curve=0,
                 variance_estimate_from="data",condition_variable="classification",
                 experimental_variable=c("experiment","plate","line"), response_variable="feature1",
                 nsimn=10, target_parameters="experiment", levels=1, max_size=15,output="test.txt")  

* power_curve = 0 : to get the power for a single level of the target parameter

* max_size = 15 : to calculate power for the case when the target parameter has 15 levels (levels=1 therefore it will test levels not sample sizes)

* output : to assign a name to the output file

Result:

Power for predictor ‘condition\_variable’, (95% confidence interval): 80.00% (44.39, 97.48) Test: Likelihood ratio Based on 10 simulations, (0 warnings, 0 errors) alpha = 0.05, nrow = 544   Time elapsed: 0 h 0 m 2 s nb: result might be an observed power calculation

Ex4. User determined effect size

Varaince estimation and power calculation from pilot data with user determined effect size of the condition variable

code example:

calculate_power(data=RMeDPower_data1,power_curve=0,
                 variance_estimate_from="data",condition_variable="classification",
                 experimental_variable=c("experiment","plate","line"), response_variable="feature1",
                 nsimn=10, target_parameters="experiment", levels=1, effect_size = c(10))  

* effect_size = 10 : to assign 10 to the effect size of condition_variable when the reesponse variable is “feature1”

Result:

Power for predictor ‘condition\_variable’, (95% confidence interval): 50.00% (18.71, 81.29) Test: Likelihood ratio Based on 10 simulations, (0 warnings, 0 errors) alpha = 0.05, nrow = 1700 Time elapsed: 0 h 0 m 10 s

Ex5. Test two target parameters

Varaince estimation and power calculation from pilot data for two target parameters: Increase the number of levels of the first target parmeter and sample sizes of the second target parameter. We will test max 9 experiments and max 142 samples per cell line. There will be two results for each parameter

code example:

calculate_power(data=RMeDPower_data1,power_curve=1,
                 variance_estimate_from="data",condition_variable="classification",
                 experimental_variable=c("experiment","plate","line"), response_variable="feature2",
                 nsimn=10, target_parameters=c("experiment","line"), levels=c(1,0), max_size=c(9,142))  

Ex6. Data with a single experimental category

If the pilot data only has a single experimental category for ‘experiment’,‘plat’ or’line’, we use known ICC values from other data.

code example:

Check sample size table of experimental varaibles:

table(RMeDPower_data2$experiment,RMeDPower_data2$plate,RMeDPower_data2$line)
#> , ,  = line1
#> 
#>       
#>         1
#>   exp1 14

calculate_power(data=RMeDPower_data2,power_curve=1,
                 variance_estimate_from="ICC",condition_variable="classification",
                 experimental_variable=c("experiment","plate","line"), response_variable="feature2",
                 nsimn=10, target_parameters=c("experiment"), levels=1, ICC=c(0.2,0.15,0.3))  

gladstone-institutes/CalcPower documentation built on Jan. 3, 2023, 11:27 a.m.