| config |
| theme |
look |
layout |
neo |
neo |
elk |
|
flowchart TD
subgraph subGraph0["Data Preparation"]
B1["Data Preprocessing"]
C["Data Quality Control"]
D["Hallmark-Specific Gene Scoring"]
end
subgraph subGraph1["Web-based tool"]
B2["Web-server"]
end
subgraph subGraph2["Gene Set Details"]
D1a["Curated Gene Sets from Databases"]
D1b["Filtered Genes with High Hazard Ratios"]
D1c["Manual Review for Biological Relevance"]
end
subgraph subGraph3["Hallmark Scoring Process"]
D1["Gene Set Curation"]
subGraph2
D2["UCell Scoring Algorithm"]
D3["Compute Digital Scores for Hallmarks"]
end
subgraph subGraph4["Thresholding and Annotation"]
E["Binary Annotation Using Otsu Thresholding"]
E1["Set Tissue-Specific Thresholds"]
E2["Classify Cells: Positive or Negative"]
end
subgraph subGraph5["Synthetic Data Creation Details"]
F1["Select 200 Cells per Patient per Hallmark"]
F2["Group by Positive and Negative Status"]
F3["Aggregate to Mimic Biopsy Samples"]
F4["Create Balanced Datasets"]
end
subgraph subGraph6["Synthetic Dataset Creation"]
F["Synthetic Bulk Dataset Creation"]
subGraph5
G["Processed Data Ready for Training"]
end
subgraph subGraph7["Model Architecture"]
H1["Shared Base Layer"]
H2["Task-Specific Layers"]
H3["Output Layers with Sigmoid Activation"]
end
subgraph subGraph8["Neural Network Model"]
H["Multi-Task Neural Network"]
subGraph7
end
subgraph subGraph9["Training Details"]
I1["Binary Cross-Entropy Loss"]
I["Training Configuration"]
I2["Adam Optimizer with Learning Rate"]
I3["Learning Rate Scheduler: Halve on No Improvement"]
I4["Early Stopping with Patience of 6 Epochs"]
end
subgraph subGraph10["Training Process"]
J1["50 Epochs with Batch Size of 256"]
J["Model Training"]
J2["Forward Pass: Compute Predictions"]
J3["Backward Pass: Compute Gradients"]
J4["Parameter Updates: Optimize Weights"]
end
subgraph subGraph11["Model Training and Validation"]
subGraph9
subGraph10
K["Validation Loss Monitoring"]
end
subgraph Metrics["Metrics"]
M1["Accuracy"]
M2["Precision"]
M3["Recall"]
M4["F1-Score"]
M5["AUROC and AUPRC"]
end
subgraph subGraph13["Testing and Evaluation"]
L["Internal Testing: 5-Fold Cross-Validation"]
M["Performance Metrics"]
Metrics
end
subgraph subGraph14["Predictive Outcomes"]
P["Predict Hallmark Activities"]
Q["Generate Detailed Report"]
R["Heatmaps, Probability Distributions, Clinical Insights"]
end
subgraph Outputs["Outputs"]
S["Sample specific analysis for personalized diagnosis and treatment"]
T["Research Insights for Oncologists"]
end
subgraph Deployment["Deployment"]
U1["Sample specific analysis"]
U2["Prediction of hallmarks of cancer"]
U3["Upload input dataset"]
U4["Web Server for User Interaction"]
end
A["Clinician/Researcher"] -- Upload single-cell Transcriptomics Data --> B1
A["Clinician/Researcher"] -- Upload Transcriptomics Data --> B2
B1 --> C
C --> D
B2 --> U4
D --> D1
D1 --> D2
D2 --> D3
D3 --> E
E --> E1
E1 --> E2
E2 --> F
F --> G
G --> H
H --> H1 & I & P
H1 --> H2
H2 --> H3
I --> I1 & I2 & I3 & I4 & J
J --> J1 & K & L
J1 --> J2 & J3 & J4
L --> M
M -- Logged --> K
K --> N["External Validation on Independent Datasets"]
N --> O["Performance Verification with Real-World Data"]
P --> Q
Q --> R
R --> S & T
U4 --> U3
U3 --> U2
U2 --> U1
G@{ shape: db}
style subGraph5 fill:#FFF9C4
style G stroke:#AA00FF,color:#000000,fill:#00C853
style subGraph7 fill:#E1BEE7
style subGraph9 fill:#E1BEE7
style subGraph10 fill:#E1BEE7
style Metrics fill:#FFF9C4
style A fill:#00C853,color:#000000
style subGraph13 fill:#FFCDD2
style subGraph3 fill:#FFCDD2
style subGraph6 fill:#FFCDD2
style subGraph8 fill:#FFE0B2
style subGraph11 fill:#FFE0B2
style subGraph0 fill:#FFCDD2
style Deployment fill:#FFE0B2
style subGraph4 fill:#FFCDD2
style subGraph14 fill:#FFE0B2
style Outputs fill:#FFCDD2
style subGraph1 fill:#FFE0B2
linkStyle 0 stroke:#000000,fill:none