The code at this page can be downloaded here

Before starting, please ensure you have installed the TriplotGUI package following the Setup instructions.

Data description

The data used in this tutorial can be downloaded here in rda format.

This example utilizes CAMP_2, which is a modified version of the CAMP data provided by the Triplot package (Schillemans et al. 2019). The CAMP dataset was simulated from authentic data collected in a cross-sectional study of carbohydrate alternatives and metabolic phenotypes among young adults in China (Liu et al. 2018). The simulated CAMP_2 data consists of three data frames:

Clinical measurements (ClinData): Contains 14 variables, including "AGE", "SEX", "BMI", "triglycerides", "total_cholesterol" "HDL"(high-density lipoprotein cholesterol), "LDL"(low-density lipoprotein cholesterol),, "GGT"(gamma-glutamyltransferase), "ALT"(alanine aminotransferase), "AST"(aspartate aminotransferase), "creatinine", "Urea_nitrogen", "Uric_acid", "Fasting_glucose".
Plasma metabolites predictive of BMI (MetaboliteData): Comprises 20 variables
Dietary intake as measured by food frequency questionnaires (FoodData): Includes 11 variables: "Refined_grains", "Coarse_grains", "Red_meat", "Poutry", "Seafood", "Egg", "Animal_organs", "Vegetables", "Fruits", "Potatos", "Legumes".

The data frames are row-wise matched by observations and consist of 300 synthetic observations.

Reference

Liu X, Liao X, Gan W, et al. Inverse Relationship between Coarse Food Grain Intake and Blood Pressure among Young Chinese Adults. Am J Hypertens. 2019;32(4):402–408. 10.1093/ajh/hpy187

Schillemans T, Shi L, Liu X, Åkesson A, Landberg R, Brunius C. Visualization and Interpretation of Multivariate Associations with Disease Risk Markers and Disease Risk-The Triplot. Metabolites. 2019 Jul 6;9(7):133. doi: 10.3390/metabo9070133

Research question

We aim to assess the relationship between diet, metabolic profiles, and risk factors for metabolic diseases, including BMI, total cholesterol, triglycerides, HDL, and LDL. In this context, we assume that metabolites act as mediators, facilitating the effect of dietary exposures on these risk factors.

Data exploration

CAMP_2 is loaded in the R environment upon running library(TriplotGUI). Let’s begin with some data exploration to understand the structure and contents of the dataset:

Check CAMP_2

Check the CAMP_2 list:

Show the code

class(CAMP_2)

[1] "list"

Show the code

names(CAMP_2)

[1] "FoodData"       "MetaboliteData" "ClinData"

Check datasets

Check the names of variables in each data:

Show the code

colnames(CAMP_2$ClinData)

 [1] "AGE"               "SEX"               "BMI"              
 [4] "triglycerides"     "total_cholesterol" "HDL"              
 [7] "LDL"               "GGT"               "ALT"              
[10] "AST"               "creatinine"        "Urea_nitrogen"    
[13] "Uric_acid"         "Fasting_glucose"

Show the code

colnames(CAMP_2$MetaboliteData)

 [1] "PC_P_20_0_22_6_"                                                                
 [2] "X1__3_4_Dihydroxyphenyl__7__4_hydroxy_3_methoxyphenyl__1_6_heptadiene_3_5_dione"
 [3] "Unknown_675.6485_9.35"                                                          
 [4] "Unknown_1066.9034_10.16"                                                        
 [5] "PC_42_8_"                                                                       
 [6] "Unknown_931.7610_9.68"                                                          
 [7] "TG_52_0_"                                                                       
 [8] "CE_22_4_"                                                                       
 [9] "Neutral_glycosphingolipids1023.67"                                              
[10] "TG_58_10_"                                                                      
[11] "Neutral_glycosphingolipids971.72"                                               
[12] "Unknown_914.7433_10.06"                                                         
[13] "Cucurbic_acid"                                                                  
[14] "DG_16_0_18_2_"                                                                  
[15] "Indoleacrylic_acid"                                                             
[16] "PC_O_20_0_22_6_"                                                                
[17] "Unknown_948.6589_8.51"                                                          
[18] "Unknown_858.5318_5.49"                                                          
[19] "O_propanoyl_carnitine"                                                          
[20] "DG_38_5_"

Show the code

colnames(CAMP_2$FoodData)

 [1] "Refined_grains" "Coarse_grains"  "Red_meat"       "Poutry"        
 [5] "Seafood"        "Egg"            "Animal_organs"  "Vegetables"    
 [9] "Fruits"         "Potatos"        "Legumes"

Check variables’ class

We then transform the data to dataframe format and use TriplotGUI’s checkdata() function to examine the classes of variables.

Show the code

ClinData<-as.data.frame(CAMP_2$ClinData)
MetaboliteData<-as.data.frame(CAMP_2$MetaboliteData)
FoodData<-as.data.frame(CAMP_2$FoodData)

ClinData_check<-checkdata(ClinData)
MetaboliteData_check<-checkdata(MetaboliteData)
FoodData_check<-checkdata(FoodData)

Show the code

ClinData_check$class_summary_statistics

$check_class_vector
              AGE               SEX               BMI     triglycerides 
        "numeric"          "factor"         "numeric"         "numeric" 
total_cholesterol               HDL               LDL               GGT 
        "numeric"         "numeric"         "numeric"         "numeric" 
              ALT               AST        creatinine     Urea_nitrogen 
        "numeric"         "numeric"         "numeric"         "numeric" 
        Uric_acid   Fasting_glucose 
        "numeric"         "numeric" 

$check_class_table
check_class_vector
 factor numeric 
      1      13

Show the code

MetaboliteData_check$class_summary_statistics

$check_class_vector
                                                                PC_P_20_0_22_6_ 
                                                                      "numeric" 
X1__3_4_Dihydroxyphenyl__7__4_hydroxy_3_methoxyphenyl__1_6_heptadiene_3_5_dione 
                                                                      "numeric" 
                                                          Unknown_675.6485_9.35 
                                                                      "numeric" 
                                                        Unknown_1066.9034_10.16 
                                                                      "numeric" 
                                                                       PC_42_8_ 
                                                                      "numeric" 
                                                          Unknown_931.7610_9.68 
                                                                      "numeric" 
                                                                       TG_52_0_ 
                                                                      "numeric" 
                                                                       CE_22_4_ 
                                                                      "numeric" 
                                              Neutral_glycosphingolipids1023.67 
                                                                      "numeric" 
                                                                      TG_58_10_ 
                                                                      "numeric" 
                                               Neutral_glycosphingolipids971.72 
                                                                      "numeric" 
                                                         Unknown_914.7433_10.06 
                                                                      "numeric" 
                                                                  Cucurbic_acid 
                                                                      "numeric" 
                                                                  DG_16_0_18_2_ 
                                                                      "numeric" 
                                                             Indoleacrylic_acid 
                                                                      "numeric" 
                                                                PC_O_20_0_22_6_ 
                                                                      "numeric" 
                                                          Unknown_948.6589_8.51 
                                                                      "numeric" 
                                                          Unknown_858.5318_5.49 
                                                                      "numeric" 
                                                          O_propanoyl_carnitine 
                                                                      "numeric" 
                                                                       DG_38_5_ 
                                                                      "numeric" 

$check_class_table
check_class_vector
numeric 
     20

Show the code

FoodData_check$class_summary_statistics

$check_class_vector
Refined_grains  Coarse_grains       Red_meat         Poutry        Seafood 
     "numeric"      "numeric"      "numeric"      "numeric"      "numeric" 
           Egg  Animal_organs     Vegetables         Fruits        Potatos 
     "numeric"      "numeric"      "numeric"      "numeric"      "numeric" 
       Legumes 
     "numeric" 

$check_class_table
check_class_vector
numeric 
     11

Check sanities for variables

To check for missing (NA) or abnormal values (e.g., NaN, negative values, infinite values, blank values) in each variable, you can use the checkdata() function. This function generates a summary table for each data frame, showing the number of observations containing NA, NaN, zero, negative, infinite (Inf), and blank values, along with their percentages.

Show the code

if (!require(kableExtra)) {
  install.packages("kableExtra")  # Install the package if it's not available.

}
library(kableExtra)             # Load it after installation.'

Show the code

kable(ClinData_check$everycolumn)

	triglycerides	GGT	ALT
NA_number	0.0000000	0.0000000	0.0000000
NA_percentage	0.0000000	0.0000000	0.0000000
NaN_number	0.0000000	0.0000000	0.0000000
NaN_percentage	0.0000000	0.0000000	0.0000000
negative_number	0.0000000	0.0000000	0.0000000
negative_percentage	0.0000000	0.0000000	0.0000000
zero_number	4.0000000	14.0000000	65.0000000
zero_percentage	0.0133333	0.0466667	0.2166667
inf_number	0.0000000	0.0000000	0.0000000
inf_percentage	0.0000000	0.0000000	0.0000000
blank_number	0.0000000	0.0000000	0.0000000
blank_percentage	0.0000000	0.0000000	0.0000000

Show the code

kable(MetaboliteData_check$everycolumn)

	Unknown_931.7610_9.68	CE_22_4_	TG_58_10_	DG_16_0_18_2_
NA_number	0.0000000	0.0000000	0.0000000	0.0000000
NA_percentage	0.0000000	0.0000000	0.0000000	0.0000000
NaN_number	0.0000000	0.0000000	0.0000000	0.0000000
NaN_percentage	0.0000000	0.0000000	0.0000000	0.0000000
negative_number	0.0000000	0.0000000	0.0000000	0.0000000
negative_percentage	0.0000000	0.0000000	0.0000000	0.0000000
zero_number	7.0000000	16.0000000	8.0000000	1.0000000
zero_percentage	0.0233333	0.0533333	0.0266667	0.0033333
inf_number	0.0000000	0.0000000	0.0000000	0.0000000
inf_percentage	0.0000000	0.0000000	0.0000000	0.0000000
blank_number	0.0000000	0.0000000	0.0000000	0.0000000
blank_percentage	0.0000000	0.0000000	0.0000000	0.0000000

Show the code

kable(FoodData_check$everycolumn)

	Refined_grains	Coarse_grains	Red_meat	Poutry	Seafood	Egg	Animal_organs	Vegetables	Fruits	Potatos	Legumes
NA_number	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.00	0.0000000	0.0000000	0.0000000
NA_percentage	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.00	0.0000000	0.0000000	0.0000000
NaN_number	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.00	0.0000000	0.0000000	0.0000000
NaN_percentage	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.00	0.0000000	0.0000000	0.0000000
negative_number	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.00	0.0000000	0.0000000	0.0000000
negative_percentage	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.00	0.0000000	0.0000000	0.0000000
zero_number	65.0000000	110.0000000	101.0000000	85.0000000	103.0000000	100.0000000	116.0000000	57.00	56.0000000	67.0000000	79.0000000
zero_percentage	0.2166667	0.3666667	0.3366667	0.2833333	0.3433333	0.3333333	0.3866667	0.19	0.1866667	0.2233333	0.2633333
inf_number	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.00	0.0000000	0.0000000	0.0000000
inf_percentage	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.00	0.0000000	0.0000000	0.0000000
blank_number	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.00	0.0000000	0.0000000	0.0000000
blank_percentage	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.0000000	0.00	0.0000000	0.0000000	0.0000000

Note

You shall only continue to use the data when the class of variables are correct and the missing or abnormal values in the variable are properly handled.

Build data for analysis

We treat food variables as our exposures and BMI, triglycerides, total_cholesterol, HDL and LDL as outcomes. We want to explore their relationships through the metabolomics data, using metabolites as assumed mediators. Sex and age are used as potential confounders for exposure-mediator and mediator-outcome association.

Show the code

exposure1 <- FoodData
Omics1 <- MetaboliteData
outcome1 <- ClinData[, c("BMI", "triglycerides", "total_cholesterol", "HDL", "LDL")]
covariates1 <- ClinData[, c("SEX" , "AGE")]

Code example of using TriplotGUI

We will use the most simple settings in this example. Please go to Tutorial(complex), if you want to try more advanced settings.

Step 1: Data reduction of Omics variables

Using TriplotGUI package, first we perform dimension reduction, i.e. principal component analysis (PCA) on metabolomics variables.

Show the code

reduced_Omics1 <- OmicsReduction(dataframe = Omics1,
                                   pcNum = 5,
                                   option = "PCA",
                                   loadingsName = TRUE)

You can view scree plot, score plot, loading plot and biplot at this stage.

Show the code

reduced_Omics1$scree_plot

plot

Show the code

reduced_Omics1$score_plot

plot

Show the code

reduced_Omics1$loading_plot

plot

Show the code

reduced_Omics1$scoreloading_plot

plot

Note

Sometimes the following text will show up in your console and the figure cannot be directly run:

libpng warning: Image width is zero in IHDR

libpng error: Invalid IHDR data

Warning message:

In dev.off(which) : agg could not write to the given file\

Run gc() in your console to clear the memory. Then rerun the code again to view the figure.

We then build a TPObject based on the data reduction results. This object will be used to save information and facilitate the transfer of data through the various steps in TriplotGUI.

Show the code

scores <- reduced_Omics1$object$scores
loadings <- reduced_Omics1$object$loadings
variance <- reduced_Omics1$object$variance
TPObject1 <- makeTPO(scores = scores,
                     loadings = loadings,
                     variance = variance)


Making TriPlotObject (TPO)
--------------------------
Loading matrix has 20 variables and 20 components.

Score matrix has 300 observations and 20 components.

TPO has 0 attached correlations.
TPO has 0 attached risks.

Step 2-3: Exposures-Omics and Omics-Outcomes associations

The associations between principal components (PCs) and food items are calculated using Pearson correlations, adjusting for confounders. We investigate associations between PC scores and risk markers adjusting for confounders using linear regression, as all food variables are numeric.

The PC scores are saved in the TPObject, which can be directly used in the exposureAssociation() function to calculate correlation coefficients and p-values with exposures (food variables), while adjusting for confounders.

Show the code

Correlations_object <- exposureAssociation(TPObject = TPObject1,
                                exposure = exposure1,
                                confounder = covariates1,
                                method = "pearson")

The correlation results are then added to the TPObject.

Show the code

TPObject2 <- addExposure(TPObject = TPObject1,
                     corrEstimate = Correlations_object$cor_estimate,
                     corrPvalue = Correlations_object$cor_pvalue)


Adding correlation to TPO
-------------------------

TPO has 11 attached correlations.
TPO has 0 attached risks.

Similarly, we calculate risk estimates (beta coefficients from linear regression) and p-values using the PC scores saved in the TPObject and the outcome variables. This is done using the outcomeAssociation() function, which also adjusts for confounders.

Show the code

Risks_object <- outcomeAssociation(TPObject = TPObject2,
                               outcome = outcome1,
                               confounder = covariates1)

The risk estimate results are then added to the TPObject.

Show the code

TPObject3 <- addOutcome(TPObject = TPObject2, 
                     Risk = Risks_object)


Adding risk to TPO
------------------

TPO has 11 attached correlations.
TPO has 5 attached risks.

Step 4: Meet-in-the-middle Triplot Visualization

Then we generate the Triplot. It’s important to note that a Triplot can be generated from any TPObject.

Show the code

TPO_plots3 <- TriplotGUI(TPObject3, 
                              riskOR = FALSE) #risks is shown as coeffcients, not odds ratio

Plot the triplot

Show the code

TPO_plots3$triplot

plot

In the triplot, black arrows represent loadings of Omics variables, forming patterns in different components (axes). Blue circle points show exposure correlations with the components, while red square points indicate outcome associations with the components.

For interpretation, let’s focus on Principal Component 1 (PC1) and Principal Component 2 (PC2), using the exposures "Refined grains", "Red meat", and the outcome "BMI" as examples:

PC1: This component shows a positive association with BMI and is correlated with high intakes of red meat and refined grains, even after adjusting for age and sex. This suggests that the metabolite features dominating PC1 may provide valuable insights into how the intake of red meat and refined grains could influence BMI.

PC2: In contrast, PC2 exhibits a smaller association with BMI and a weak inverse correlation with red meat and refined grains. The metabolite features contributing significantly to PC2 may indicate a marginal beneficial effect of these dietary components, although this effect appears to be limited. This may also suggest that the features dominating PC2 may reflect metabolite patterns influenced by food, but not to a large degree associations with metabolic regulation in relation to health.

This interpretation highlights how different dietary exposures relate to metabolic outcomes, offering insights into potential mechanisms at play.

Step 5: Mediation analysis and visualization

Mediation analysis will be performed using the conventional(Baron-Kenny) method on our selected exposures ("Refined grains", "Red meat"), and the outcome of interest ("BMI"), using components generated from Step 1 as mediators and adjusting for age and sex in both the exposure-mediator and mediator-outcome associations.

Show the code

mediation_object3 <- 
  getMediationConventional (mediator = TPObject3$scores[,c(1:2),drop = FALSE], 
                             # Specfiying at least 2 components so that there can be a 2-dimensional plot
                             exposure = exposure1[,c("Refined_grains","Red_meat"),drop = FALSE],
                             outcome = outcome1[,"BMI",drop = FALSE],
                             confounderME = covariates1,
                             confounderOE = covariates1)

After mediation analysis is performed, users can select which exposure, mediator and outcome variables to use for mediation barplot visualization. Users can examine the barplot for a clearer view of the direct, indirect, and total effects for each combination of PC, exposure, and outcome.

Show the code

mediation_plots3 <- mediationBarplot(mediationObject = mediation_object3,
                                         cex = 2,
                                         # size of the text
                                         by_row = "one_column")

Show the code

mediation_plots3

plot

Note

The barplot displays direct, indirect, and total effects for each exposure-mediator-outcome combination, providing a convenient tool to assess the direction and magnitude of mediation estimates.

Color coding indicates significance and direction: red represents significant positive effects (p<0.05), blue represents significant negative effects (p<0.05), and grey represents insignificant effects.
Significance levels are denoted by stars: one star for p<0.05, two stars for p<0.01, and three stars for p<0.001.

The mediation results are then added to the TPObject.

Show the code

TPObject4 <- addMediation(TPObject = TPObject3, 
                           mediationObject = mediation_object3)


Adding mediation to TPO
-------------------------

TPO mediation has used 2 mediators, 2 exposures, 1 outcomes
Mediators are PC1, PC2 
Exposures are Refined_grains, Red_meat 
Outcomes are BMI

Based on the mediation barplot, we observed significant mediation effects through PC1 for the red_meat-BMI and refined_grain-BMI associations. This finding suggests that the metabolite features contributing to PC1 are likely mediating the pathway from red meat and refined grain intake to BMI changes. These results provide insight into potential metabolic mechanisms linking dietary patterns to body mass index.

Comparative Visualization

Users can easily visualize the exposure associations, risk estimations, and mediation results stored in the TPObject using the checkTPO() function provided by TriplotGUI. This function generates heatmaps for each type of analysis, offering a comprehensive and intuitive way to examine the relationships between variables. Note that you could put any TPObjects generate along the steps for heatmap visualizations.

Show the code

checkTPObject4 <- checkTPO(TPObject4,
                           mediators=c("PC1","PC2"),
                                  size=7    ## customize the size of stars
                                  )

Show heatmap for correlation coefficients and p values between PCs and exposures:

Show the code

checkTPObject4$corr_coefficients

plot

Show heatmap for risk estimates (beta coefficients) and p values between PCs and outcomes:

Show the code

checkTPObject4$risk_coefficients

plot

Show heatmap for mediation estimates and p values for indirect effect(IE), and total effect(TE), proportion mediated (PM) and adjusted proportion mediated (APM): See Manual section for more details:

Show the code

checkTPObject4$med_coefficients

plot

Note

Significance is indicated as follows. One star for p<0.05; Two stars for p<0.01; Three stars for p<0.001.
Note that in the heatmap from checkTPbject4$med_coefficients, only the selected exposures, mediators and outcomes that we use to do mediation on will show up. The rows in the heatmaps are mediators and the column represents exposure-outcome pairs. For each exposure-outcome pair, 4 result are shown: IE (indirect effect), TE (total effect), PM(Proportion Mediated) and APM(Adjusted Proportion Mediated). Please see Details section for more information. The “BK” before the names of exposures-outcome pairs means that this mediation is using the Coventional Baron-Kenny “product” method. For this method, only the significant level of IE and TE is calculated and shown on the plot, but not PM and APM.

Data Download

To save all your output, including data, results, and visualization output as an .rda file, you can use the save() function in R.

Show the code

save(exposure1,Omics1,outcome1,covariates1,
     reduced_Omics1,Correlations_object,Risks_object,
     mediation_object3,mediation_plots3,
     TPObject1,TPObject2,TPObject3,TPObject4,
     checkTPObject4,
     "Tutorial_simple_output.rda")

--- title: "Tutorial(simple) - Code" filters: - lightbox lightbox: auto subtitle: "Simple demo of using TriplotGUI's code" title-block-banner: true title-block-banner-color: white title-block-categories: false page-layout: full format: html: margin-top: 0em margin-bottom: 0em padding-top: 0em padding-bottom: 0em minimal: true smooth-scroll: true fig-responsive: true toc-location: right toc-depth: 5 toc-title: Tutorial #number-sections: true ## add number in the session or not #number-depth: 4 code-fold: show code-summary: "Show the code" code-overflow: wrap code-tools: true code-copy: true highlight: tango df-print: paged ## kable is another option How the dataframe looks like standalone: false ### specifies if all assets and libraries must be integrated into the output html file as a standalone document. fig-align: right theme: solar geometry: margin=lin --- ```{r, include=FALSE} ##This is a function that combines html and rmarkdown. It aims to generate a button that could hide output or plot (fold.output=T or fold.plot=T). hooks = knitr::knit_hooks$get() hook_foldable = function(type) { force(type) function(x, options) { res = hooks[[type]](x, options) if (!isTRUE(options[[paste0("fold.", type)]])) return(res) paste0( "<details><summary>", type, "</summary>\n\n", res, "\n\n</details>" ) } } knitr::knit_hooks$set( output = hook_foldable("output"), plot = hook_foldable("plot") ) ``` The code at this page can be downloaded [here](https://gitlab.com/YingxiaoYan/triplotgui/-/blob/main/Tutorial_code/CAMP_2_applying_TriplotGUI_code.R){target="_blank"} Before starting, please ensure you have installed the TriplotGUI package following the [Setup](installation.qmd){target="_blank" style="color:blue"} instructions. # Data description The data used in this tutorial can be downloaded [here](https://gitlab.com/YingxiaoYan/triplotgui/-/blob/main/data/CAMP_2.rda){target="_blank"} in rda format. This example utilizes ***CAMP_2***, which is a modified version of the **CAMP** data provided by the [*Triplot*](https://gitlab.com/CarlBrunius/triplot/-/blob/master/vignettes/Triplot_Tutorial.md?ref_type=heads){target="_blank"} package (Schillemans et al. 2019). The **CAMP** dataset was simulated from authentic data collected in a cross-sectional study of carbohydrate alternatives and metabolic phenotypes among young adults in China (Liu et al. 2018). The simulated ***CAMP_2*** data consists of three data frames: - Clinical measurements (ClinData): Contains 14 variables, including `"AGE"`, `"SEX"`, `"BMI"`, `"triglycerides"`, `"total_cholesterol"` `"HDL"`(high-density lipoprotein cholesterol), `"LDL"`(low-density lipoprotein cholesterol),, `"GGT"`(gamma-glutamyltransferase), `"ALT"`(alanine aminotransferase), `"AST"`(aspartate aminotransferase), `"creatinine"`, `"Urea_nitrogen"`, `"Uric_acid"`, `"Fasting_glucose"`. - Plasma metabolites predictive of BMI (MetaboliteData): Comprises 20 variables - Dietary intake as measured by food frequency questionnaires (FoodData): Includes 11 variables: `"Refined_grains"`, `"Coarse_grains"`, `"Red_meat"`, `"Poutry"`, `"Seafood"`, `"Egg"`, `"Animal_organs"`, `"Vegetables"`, `"Fruits"`, `"Potatos"`, `"Legumes"`. The data frames are row-wise matched by observations and consist of 300 synthetic observations. **Reference** Liu X, Liao X, Gan W, et al. Inverse Relationship between Coarse Food Grain Intake and Blood Pressure among Young Chinese Adults. *Am J Hypertens*. 2019;32(4):402–408. [10.1093/ajh/hpy187](https://academic.oup.com/ajh/article/32/4/402/5236579?login=true){target="_blank"} Schillemans T, Shi L, Liu X, Åkesson A, Landberg R, Brunius C. Visualization and Interpretation of Multivariate Associations with Disease Risk Markers and Disease Risk-The Triplot. *Metabolites*. 2019 Jul 6;9(7):133. [doi: 10.3390/metabo9070133](https://www.mdpi.com/2218-1989/9/7/133){target="_blank"} # Research question We aim to assess the relationship between diet, metabolic profiles, and risk factors for metabolic diseases, including BMI, total cholesterol, triglycerides, HDL, and LDL. In this context, we assume that metabolites act as mediators, facilitating the effect of dietary exposures on these risk factors. # Data exploration ***CAMP_2*** is loaded in the R environment upon running `library(TriplotGUI)`. Let's begin with some data exploration to understand the structure and contents of the dataset: ```{r noshowcode_runcode_noshowoutput_1} #| eval: true #| echo: false #| output: false #| classes: styled-output library(TriplotGUI) ``` ## Check CAMP_2 Check the ***CAMP_2*** list: ```{r showcode_runcode_showoutput_1} #| eval: true #| echo: true #| output: true #| classes: styled-output class(CAMP_2) names(CAMP_2) ``` ## Check datasets Check the names of variables in each data: ```{r showcode_runcode_showoutput_2} #| eval: true #| echo: true #| output: true #| classes: styled-output colnames(CAMP_2$ClinData) colnames(CAMP_2$MetaboliteData) colnames(CAMP_2$FoodData) ``` ## Check variables' class We then transform the data to dataframe format and use TriplotGUI's `checkdata()` function to examine the classes of variables. ```{r showcode_runcode_noshowoutput_1} #| eval: true #| echo: true #| output: false #| classes: styled-output ClinData<-as.data.frame(CAMP_2$ClinData) MetaboliteData<-as.data.frame(CAMP_2$MetaboliteData) FoodData<-as.data.frame(CAMP_2$FoodData) ClinData_check<-checkdata(ClinData) MetaboliteData_check<-checkdata(MetaboliteData) FoodData_check<-checkdata(FoodData) ``` ```{r showcode_runcode_showoutput_3} #| eval: true #| echo: true #| output: true #| classes: styled-output ClinData_check$class_summary_statistics MetaboliteData_check$class_summary_statistics FoodData_check$class_summary_statistics ``` ## Check sanities for variables To check for missing (NA) or abnormal values (e.g., NaN, negative values, infinite values, blank values) in each variable, you can use the `checkdata()` function. This function generates a summary table for each data frame, showing the number of observations containing NA, NaN, zero, negative, infinite (Inf), and blank values, along with their percentages. ```{r showcode_runcode_noshowoutput_1_1} #| eval: true #| echo: true #| output: false #| classes: styled-output if (!require(kableExtra)) { install.packages("kableExtra") # Install the package if it's not available. } library(kableExtra) # Load it after installation.' ``` ```{r showcode_runcode_showoutput_4} #| eval: true #| echo: true #| output: true #| classes: styled-output kable(ClinData_check$everycolumn) kable(MetaboliteData_check$everycolumn) kable(FoodData_check$everycolumn) ``` ::: callout-note You shall only continue to use the data when the class of variables are correct and the missing or abnormal values in the variable are properly handled. ::: ## Build data for analysis We treat food variables as our exposures and BMI, triglycerides, total_cholesterol, HDL and LDL as outcomes. We want to explore their relationships through the metabolomics data, using metabolites as assumed mediators. Sex and age are used as potential confounders for exposure-mediator and mediator-outcome association. ```{r showcode_runcode_showoutput_5} #| eval: true #| echo: true #| output: true #| classes: styled-output exposure1 <- FoodData Omics1 <- MetaboliteData outcome1 <- ClinData[, c("BMI", "triglycerides", "total_cholesterol", "HDL", "LDL")] covariates1 <- ClinData[, c("SEX" , "AGE")] ``` # Code example of using TriplotGUI We will use the most simple settings in this example. Please go to [Tutorial(complex)](complex_code.qmd){target="_blank"}, if you want to try more advanced settings. ## Step 1: Data reduction of Omics variables {#simple_link} Using `TriplotGUI` package, first we perform dimension reduction, i.e. principal component analysis (PCA) on metabolomics variables. ```{r showcode_runcode_noshowoutput_2} #| eval: true #| echo: true #| output: false #| classes: styled-output reduced_Omics1 <- OmicsReduction(dataframe = Omics1, pcNum = 5, option = "PCA", loadingsName = TRUE) ``` You can view scree plot, score plot, loading plot and biplot at this stage. ```{r showcode_runcode_showoutput_6} #| eval: true #| echo: true #| output: true #| classes: styled-output #| fold.plot: true #| out-width: "100%" reduced_Omics1$scree_plot reduced_Omics1$score_plot reduced_Omics1$loading_plot reduced_Omics1$scoreloading_plot ``` ::: callout-note Sometimes the following text will show up in your console and the figure cannot be directly run: `libpng warning: Image width is zero in IHDR` `libpng error: Invalid IHDR data` `Warning message:` `In dev.off(which) : agg could not write to the given file\` Run `gc()` in your console to clear the memory. Then rerun the code again to view the figure. ::: We then build a TPObject based on the data reduction results. This object will be used to save information and facilitate the transfer of data through the various steps in TriplotGUI. ```{r showcode_runcode_showoutput_7} #| eval: true #| echo: true #| output: true #| classes: styled-output #| fold.output: false scores <- reduced_Omics1$object$scores loadings <- reduced_Omics1$object$loadings variance <- reduced_Omics1$object$variance TPObject1 <- makeTPO(scores = scores, loadings = loadings, variance = variance) ``` ## Step 2-3: Exposures-Omics and Omics-Outcomes associations The associations between principal components (PCs) and food items are calculated using Pearson correlations, adjusting for confounders. We investigate associations between PC scores and risk markers adjusting for confounders using linear regression, as all food variables are numeric. The PC scores are saved in the TPObject, which can be directly used in the `exposureAssociation()` function to calculate correlation coefficients and p-values with exposures (food variables), while adjusting for confounders. ```{r showcode_runcode_noshowoutput_3} #| eval: true #| echo: true #| output: false #| classes: styled-output #| fold.output: false Correlations_object <- exposureAssociation(TPObject = TPObject1, exposure = exposure1, confounder = covariates1, method = "pearson") ``` The correlation results are then added to the TPObject. ```{r showcode_runcode_showoutput_8} #| eval: true #| echo: true #| output: true #| classes: styled-output #| fold.output: false TPObject2 <- addExposure(TPObject = TPObject1, corrEstimate = Correlations_object$cor_estimate, corrPvalue = Correlations_object$cor_pvalue) ``` Similarly, we calculate risk estimates (beta coefficients from linear regression) and p-values using the PC scores saved in the TPObject and the outcome variables. This is done using the `outcomeAssociation()` function, which also adjusts for confounders. ```{r showcode_runcode_noshowoutput_4} #| eval: true #| echo: true #| output: false #| classes: styled-output #| fold.output: false Risks_object <- outcomeAssociation(TPObject = TPObject2, outcome = outcome1, confounder = covariates1) ``` The risk estimate results are then added to the TPObject. ```{r showcode_runcode_showoutput_9} #| eval: true #| echo: true #| output: true #| classes: styled-output #| fold.output: false TPObject3 <- addOutcome(TPObject = TPObject2, Risk = Risks_object) ``` ## Step 4: Meet-in-the-middle Triplot Visualization Then we generate the Triplot. It's important to note that a Triplot can be generated from any TPObject. ```{r showcode_runcode_noshowoutput_5} #| eval: true #| echo: true #| output: false #| classes: styled-output #| fold.output: false TPO_plots3 <- TriplotGUI(TPObject3, riskOR = FALSE) #risks is shown as coeffcients, not odds ratio ``` Plot the triplot ```{r showcode_runcode_showoutput_10} #| eval: true #| echo: true #| output: true #| classes: styled-output #| fold.plot: true #| out-width: "100%" TPO_plots3$triplot ``` In the triplot, black arrows represent loadings of Omics variables, forming patterns in different components (axes). Blue circle points show exposure correlations with the components, while red square points indicate outcome associations with the components. For interpretation, let’s focus on Principal Component 1 (PC1) and Principal Component 2 (PC2), using the exposures `"Refined grains"`, `"Red meat"`, and the outcome `"BMI"` as examples: PC1: This component shows a positive association with BMI and is correlated with high intakes of red meat and refined grains, even after adjusting for age and sex. This suggests that the metabolite features dominating PC1 may provide valuable insights into how the intake of red meat and refined grains could influence BMI. PC2: In contrast, PC2 exhibits a smaller association with BMI and a weak inverse correlation with red meat and refined grains. The metabolite features contributing significantly to PC2 may indicate a marginal beneficial effect of these dietary components, although this effect appears to be limited. This may also suggest that the features dominating PC2 may reflect metabolite patterns influenced by food, but not to a large degree associations with metabolic regulation in relation to health. This interpretation highlights how different dietary exposures relate to metabolic outcomes, offering insights into potential mechanisms at play. ## Step 5: Mediation analysis and visualization Mediation analysis will be performed using the conventional(Baron-Kenny) method on our selected exposures (`"Refined grains"`, `"Red meat"`), and the outcome of interest (`"BMI"`), using components generated from Step 1 as mediators and adjusting for age and sex in both the exposure-mediator and mediator-outcome associations. ```{r showcode_runcode_noshowoutput_6} #| eval: true #| echo: true #| output: false #| classes: styled-output #| fold.output: false mediation_object3 <- getMediationConventional (mediator = TPObject3$scores[,c(1:2),drop = FALSE], # Specfiying at least 2 components so that there can be a 2-dimensional plot exposure = exposure1[,c("Refined_grains","Red_meat"),drop = FALSE], outcome = outcome1[,"BMI",drop = FALSE], confounderME = covariates1, confounderOE = covariates1) ``` After mediation analysis is performed, users can select which exposure, mediator and outcome variables to use for mediation barplot visualization. Users can examine the barplot for a clearer view of the direct, indirect, and total effects for each combination of PC, exposure, and outcome. ```{r showcode_runcode_noshowoutput_7} #| eval: true #| echo: true #| output: false #| classes: styled-output #| fold.plot: true mediation_plots3 <- mediationBarplot(mediationObject = mediation_object3, cex = 2, # size of the text by_row = "one_column") ``` ```{r showcode_runcode_showoutput_11} #| eval: true #| echo: true #| output: true #| classes: styled-output #| fold.plot: true #| out-width: "100%" #| out-height: "200%" mediation_plots3 ``` ::: callout-note The barplot displays direct, indirect, and total effects for each exposure-mediator-outcome combination, providing a convenient tool to assess the direction and magnitude of mediation estimates. - Color coding indicates significance and direction: red represents significant positive effects (p\<0.05), blue represents significant negative effects (p\<0.05), and grey represents insignificant effects. - Significance levels are denoted by stars: one star for p\<0.05, two stars for p\<0.01, and three stars for p\<0.001. ::: The mediation results are then added to the TPObject. ```{r showcode_runcode_showoutput_12} #| eval: true #| echo: true #| output: true #| classes: styled-output #| fold.plot: true TPObject4 <- addMediation(TPObject = TPObject3, mediationObject = mediation_object3) ```                                   Based on the mediation barplot, we observed significant mediation effects through PC1 for the red_meat-BMI and refined_grain-BMI associations. This finding suggests that the metabolite features contributing to PC1 are likely mediating the pathway from red meat and refined grain intake to BMI changes. These results provide insight into potential metabolic mechanisms linking dietary patterns to body mass index. ## Comparative Visualization Users can easily visualize the exposure associations, risk estimations, and mediation results stored in the TPObject using the `checkTPO()` function provided by TriplotGUI. This function generates heatmaps for each type of analysis, offering a comprehensive and intuitive way to examine the relationships between variables. Note that you could put any TPObjects generate along the steps for heatmap visualizations. ```{r showcode_runcode_noshowoutput_9} #| eval: true #| echo: true #| output: false #| classes: styled-output #| fold.plot: true checkTPObject4 <- checkTPO(TPObject4, mediators=c("PC1","PC2"), size=7 ## customize the size of stars ) ``` Show heatmap for correlation coefficients and p values between PCs and exposures: ```{r showcode_runcode_showoutput_15A} #| eval: true #| echo: true #| output: true #| classes: styled-output #| fold.plot: true #| out-width: "100%" checkTPObject4$corr_coefficients ``` Show heatmap for risk estimates (beta coefficients) and p values between PCs and outcomes: ```{r showcode_runcode_showoutput_15B} #| eval: true #| echo: true #| output: true #| classes: styled-output #| fold.plot: true checkTPObject4$risk_coefficients ``` Show heatmap for mediation estimates and p values for indirect effect(IE), and total effect(TE), proportion mediated (PM) and adjusted proportion mediated (APM): See [Manual](workflow.qmd){target="_blank"} section for more details: ```{r showcode_runcode_showoutput_15C} #| eval: true #| echo: true #| output: true #| classes: styled-output #| fold.plot: true #| out-width: "100%" checkTPObject4$med_coefficients ``` ::: callout-note - Significance is indicated as follows. One star for p\<0.05; Two stars for p\<0.01; Three stars for p\<0.001. - Note that in the heatmap from `checkTPbject4$med_coefficients`, only the selected exposures, mediators and outcomes that we use to do mediation on will show up. The rows in the heatmaps are mediators and the column represents exposure-outcome pairs. For each exposure-outcome pair, 4 result are shown: *IE* (indirect effect), *TE* (total effect), *PM*(Proportion Mediated) and *APM*(Adjusted Proportion Mediated). Please see [Details section](https://yingxiaoyan.gitlab.io/triplotgui_tutorial/workflow.html#5.2){target="_blank"} for more information. The "BK" before the names of exposures-outcome pairs means that this mediation is using the Coventional Baron-Kenny "product" method. For this method, only the significant level of *IE* and *TE* is calculated and shown on the plot, but not *PM* and *APM*. ::: ## Data Download To save all your output, including data, results, and visualization output as an .rda file, you can use the `save()` function in R. ```{r showcode_noruncode_noshowoutput_1} #| eval: false #| echo: true #| warning: false #| error: true #| classes: styled-output save(exposure1,Omics1,outcome1,covariates1, reduced_Omics1,Correlations_object,Risks_object, mediation_object3,mediation_plots3, TPObject1,TPObject2,TPObject3,TPObject4, checkTPObject4, "Tutorial_simple_output.rda") ```