ADAM
Advanced Design & AI-Driven Modeling for Plant Tissue Culture Media Optimization

Experiment Design Dashboard

Design Type
Total Runs
Factors
Design Generated: Download your experimental design and conduct laboratory experiments.
Ready to Design: Configure your factors and generate an experimental design.

No Design Generated

Configure your factors and generate a design to see the experimental design matrix.

Experimental Design Matrix


                      

                    

                  

                

              

            

          

          

            

              

                
              

              

                

                  
Data Analysis Dashboard

                  

                    

                      

                        
Dataset Status

                        

                          

                        

                      

                    

                    

                      

                        
Total Rows

                        

                          
                        

                      

                    

                    

                      

                        
Predictors

                        

                          
                        

                      

                    

                    

                      

                        
Responses

                        

                          

                        

                      

                    

                  

                  

                    
Data Ready: Proceed to Step 3 to build machine learning models.

                  

                  

                    
Upload Required: Upload your experimental results CSV file.

                  

                

                

                  
                  

                    

                      

                        

                          
No Data Uploaded

                          
Please upload a CSV file to begin.

                        

                      

                      

                        

                          
Data Preview

                          

                        

                      

                    

                    

                      

                        

                        

                        
Data subsetting

                        

                          Recommended: Use Descriptor variables (Type A) for subsetting

                          Descriptors: Treatment IDs, Genotypes, Runs, Blocks, etc.
                        

                        

                          

                            

                            

                              
                              
                            

                          

                        

                        

                          

                            
                          

                        

                      

                    

                    

                      

                        
Data Quality Assessment

                        
Analyze your data for missing values, outliers, correlations, and data types

                        

                          

                            

                              

                                
Missing Values

                                

                              

                            

                            

                              

                                
Outliers

                                

                              

                            

                            

                              

                                
Correlations

                                

                              

                            

                            

                              

                                
Data Types

                                

                              

                            

                          

                          

                            
                          

                          

                          

                            
Missing Values

                            

                            

                          

                          

                            
Outliers Detected

                            

                            

                          

                          

                            
High Correlations

                            

                            

                          

                        

                      

                    

                    

                      

                        
Data Distribution & Relationships

                        

                          
                          

                            

                              

                            

                            

                              

                            

                            

                              

                                
Predictor-Response Relationship Explorer

                                

                                  

                                    

                                      
                                      

                                        
                                        
                                      

                                    

                                  

                                  

                                    

                                      
                                      

                                        
                                        
                                      

                                    

                                  

                                  

                                    

                                      
                                    

                                  

                                

                                

                              

                            

                            

                              

                            

                            

                              

                            

                          

                        

                      

                    

                  

                

              

            

          

          

            

              

                
              

              

                

                  
Machine Learning Dashboard

                  

                    

                      

                        
ML Status

                        

                          

                        

                      

                    

                    

                      

                        
Trained Models

                        

                          
                        

                      

                    

                    

                      

                        
Best Algorithm

                        

                          
                        

                      

                    

                    

                      

                        
Best Test R²

                        

                          
                        

                      

                    

                  

                  

                    
Models Ready: Proceed to Step 4 to optimize your experimental conditions.

                  

                  

                    
Ready for Training: Configure settings and train machine learning models.

                  

                  

                    
Prerequisites Missing: Complete data upload and variable classification first.

                  

                

                

                  
                  

                    

                      

                        

                          
ML Dataset Overview

                          

                            

                              

                                

                                  
Predictors

                                  

                                    
                                  

                                

                                

                                  
Responses

                                  

                                    
                                  

                                

                              

                            

                            

                              

                                

                              

                            

                          

                          
Test Set Split Preview

                          

                          
Variable Summary

                          

                        

                      

                      

                        

                          
ML Dataset Not Available

                          
Please upload data and complete variable classification to view the ML dataset.

                        

                      

                    

                    

                      

                        

                          
Performance Comparison

                          

                          
                            
                            Download Data
                          
                        

                        

                          
Model Comparison

                          

                        

                      

                      

                        

                          No spot check results available.

                          Click 'Spot Check All Models' to train and compare all available algorithms.
                        

                      

                    

                    

                      

                        
Model Metrics

                        

                          
                          

                            
                            
                          

                        

                        

                          
Prediction Performance

                          

                          
                            
                            Download Data
                          
                        

                        

                        

                          
Variable Importance

                          

                          
                            
                            Download Data
                          
                        

                      

                    

                    

                      

                        

                          
                          Getting Started: Configure and train models in the settings panel above.
                        

                      

                      

                        
Model Repository

                        

                          

                            
                          

                          

                            
                          

                          

                        

                        

                          

                        

                      

                      

                        
Ensemble Model Training

                        

                          

                            

                              
                              

                                
                                
                              

                            

                          

                          

                            

                              
                              

                                
                                
                              

                            

                          

                        

                        

                          

                            

                              
Simple Models

                              

                              

                                
                                

                              

                            

                          

                          

                            

                              
Tree-based Models

                              

                              

                                
                                

                              

                            

                          

                          

                            

                              
Complex Models

                              

                              

                                
                                

                              

                            

                          

                        

                        

                        

                          

                            

                              
                              
                            

                          

                          

                            

                              

                                
                              

                              

                                

                                  
                                  Ensemble Training...
                                

                              

                            

                          

                        

                        

                          Status: 
                          
                        

                        

                          

                            Scaling Warning: 
                            
                          

                        

                        

                          

                            Ensemble training combines predictions from multiple base models using a meta-learner (default: Linear Regression).
                             The ensemble uses the same validation settings (CV method, folds, repeats) configured in the sidebar.
                            

                            Important: 
                            All selected models must have the same scaling status and target the same response variable.
                          

                        

                      

                    

                  

                

              

            

          

          

            

              

                
              

              

                

                  
Optimization Dashboard

                  

                    

                      

                        
Status

                        

                          

                        

                      

                    

                    

                      

                        
Last Algorithm

                        

                          
                        

                      

                    

                    

                      

                        
Initial ↔ Final Best

                        

                          

                        

                      

                    

                    

                      

                        
Found Solutions

                        

                          
                        

                      

                    

                  

                  

                    
Optimization Complete: Proceed to Step 5 to evaluate solutions.

                  

                  

                    
Prerequisites Missing: Complete data analysis and model training first.

                  

                

                

                  
                  

                    

                      

                        

                          
No Optimization Results Available

                          
Run an optimization to view fitness evolution and progress metrics.

                        

                      

                      

                        

                          
Fitness Evolution

                          

                          
                            
                            Download Data
                          
                        

                        

                          

                            
Individual Response Variables

                            

                            
                              
                              Download Data
                            
                          

                        

                        

                          
Detailed Status

                          

                        

                      

                    

                    

                      

                        
Complete Optimization Results

                        

                          Filter Status: 
                          
                        

                        

                          
Solution Filtering

                          

                            

                          

                        

                        

                        

                      

                    

                    

                      

                        
Parameter Space Exploration

                        
Visualize how the algorithm explored the parameter space during optimization. Colors indicate solution quality.

                        

                        

                        
                          
                          Download Data
                        
                        
Parameter Space Guide:

                          • Single-objective: Color = fitness value (red=low, green=high)

                          • Multi-objective: Color = proximity to ideal solution

                          • Density contours show exploration patterns

                      

                      

                        
Objective Space Visualization (Pareto Front)

                        
Visualize solutions in objective/response space. For multi-objective optimization, the Pareto front is highlighted.

                        

                        

                        
                          
                          Download Data
                        
                        

                          
                          Objective Space Guide:

                          • Multi-objective: Red dashed line = Pareto front

                          • Colors: Points colored by fitness (if available)
                        

                      

                    

                    

                      

                        
Solution Analysis

                        
Solution Diversity

                        
Distribution of pairwise distances showing solution spread in parameter space.

                        

                        
                          
                          Download Data
                        
                        

                        
Performance Metrics Summary

                        

                      

                    

                    

                      

                        
Optimization History

                        
Track and compare multiple optimization runs to understand algorithm performance patterns.

                        

                          
                        

                        

                        

                        
Algorithm Performance Comparison

                        
Compare runtime vs. solution quality across different algorithms and parameter settings.

                        

                        
                          
                          Download Data
                        
                      

                    

                  

                

              

            

          

          

            

              

                
              

              

                

                  
Solution Validation Dashboard

                  

                    

                      

                        
Validation Status

                        

                          

                        

                      

                    

                    

                      

                        
Response Variables

                        

                          

                        

                      

                    

                    

                      

                        
Best Improvement

                        

                          

                        

                      

                    

                    

                      

                        
Avg. Improvement

                        

                          

                        

                      

                    

                  

                  

                    
Validation Complete: Review comparison table and improvement metrics below.

                  

                  

                    
Prerequisites Missing: Complete optimization in Step 4 first.

                  

                

                

                  

                  

                    
No Validation Results Available

                    
Run an optimization see solution comparison.

                  

                

                

                  

                    
Distribution Analysis

                    

                    
                  

                  

                    
Dataset Comparison

                    

                    
                  

                

              

            

          

          

            

              

                

                  

                    

                    

                      

                        

                          

                            
                              
                                
                                  
                                  
                                
                              
                              
                              
                              
                              
                              
                              
                              
                                
                                
                                
                              
                              
                                
                                
                                
                                
                              
                              
                                
                                
                                
                              
                            
                            

                              
ADAM

                              
Advanced Design & AI-Driven Modeling for Plant Tissue Culture Media Optimization

                            

                          

                        

                        
This application helps researchers design and optimize experiments for plant tissue culture media using machine learning and evolutionary algorithms.

                        

                          
How to use ADAM:

                          

                            

                              1
                            

                            

                              
Step 1 : Design experiments

                              
Generate a statistically sound design of experiments.

                            

                          

                          

                            

                              2
                            

                            

                              
Step 2 : Analyze data

                              
Upload your results to verify data quality and identify patterns.

                            

                          

                          

                            

                              3
                            

                            

                              
Step 3 : Train predictive models

                              
Build machine learning models to understand relationships.

                            

                          

                          

                            

                              4
                            

                            

                              
Step 4 : Optimize media compositions

                              
Use evolutionary algorithms to find optimal factor combinations.

                            

                          

                          

                            

                              5
                            

                            

                              
Step 5 : Evaluate solutions

                              
Compare found solutions to uploaded data.

                            

                          

                          

                          

                            

                              

                                
Latest Features

                                

                                  

                                    Categorical predictors
                                     - Native categorical variable support
                                  

                                  

                                    Solution filtering
                                     - Distance-based similarity filtering
                                  

                                  

                                    Optimization constraints
                                     - Equality and inequality constraints
                                  

                                  

                                    LHS Implementation
                                     - Latin Hypercube Sampling
                                  

                                  

                                    Enhanced visualizations
                                     - Improved data quality plots
                                  

                                  

                                    New appearance
                                     - Layout updates and bug fixes
                                  

                                

                              

                            

                            

                              

                                
Upcoming Features

                                

                                  

                                    Sensitivity analysis
                                     - Global and local sensitivity with Sobol indices
                                  

                                  

                                    Bayesian optimization
                                     - GP-based surrogate modeling
                                  

                                

                              

                            

                          

                        

                      

                      

                        

                          

                          

                            
Citation

                            
Hans Bethge (2025). ADAM - AI-assisted optimization. Version V2.1.6

                          

                          

                          
                          

                          

                            
Funding

                            
This project was funded by E-Cost action CopyTree.

                          

                          

                          

                            
Main Libraries

                            

                              

                                • 
                                shiny
                                 - Web Application Framework
                              

                              
Chang, W., et al. (2024). V1.10.0

                              

                                • 
                                caret
                                 - Machine Learning Framework
                              

                              
Kuhn, M. (2008). V7.0.1

                              

                                • 
                                ecr
                                 - Evolutionary Computation
                              

                              
Bossek, J. (2017). V2.1.1

                              

                              

                                Developed with assistance from 
                                Claude AI
                                .
                              

                            

                          

                          

                          

                            
💬 Feedback & Contact

                            
Have ideas or suggestions? We'd love to hear from you!

                            
                          

                        

                      

                    

                  

                

              

            

          

          

            

              

                

                  
System Information