Add GTEx Image Learner tutorial for tissue classification using expression-derived images

[paulocilasjr]

Title: Add GTEx Image Learner tutorial for tissue classification using expression-derived images

---

Description:

This PR introduces a new tutorial demonstrating how to use the Galaxy Image Learner to perform tissue classification using GTEx v11 gene expression data transformed into images.

The tutorial provides a complete, end-to-end workflow, guiding users from prepared input data to model training, evaluation, and interpretation of results.

---

✨ Key Features of the Tutorial

1. End-to-end workflow

• Uses preprocessed GTEx v11 expression-derived images and metadata
• Trains an Image Learner model for multi-class tissue classification
• Guides users through dataset selection, configuration, training, and evaluation

---

2. Integration of biological data with deep learning

• Demonstrates how high-dimensional gene expression data can be transformed into image representations
• Shows how deep learning models can capture tissue-specific expression patterns
• Provides a practical example of combining bioinformatics and machine learning within Galaxy

---

3. Model evaluation and interpretation

• Includes detailed guidance on interpreting:
  • Training, validation, and test metrics
  • Accuracy, ROC-AUC, loss, and Hits@K
• Explains how to analyze training behavior (learning curves, overfitting)
• Introduces Grad-CAM visualizations for model interpretability

---

4. Emphasis on responsible interpretation

• Highlights that near-perfect performance should be interpreted carefully
• Discusses limitations of the expression-to-image transformation
• Encourages validation on additional datasets and real-world scenarios

---

5. Learner-focused design

• Step-by-step instructions with clear explanations
• Figures illustrating key steps (e.g., accessing the report)
• Structured sections that balance usability and conceptual understanding

---

📌 Notes

• The tutorial is designed as both a teaching example and a benchmark workflow for the Image Learner tool.
• It does not assume that gene expression data is inherently image-like, but rather uses this representation as a modeling approach.

✅ Checklist

• Descriptive title
• Detailed description of tutorial scope and content
• TODO items listed
• Images reviewed for GTN compatibility

---

This tutorial provides a practical and educational example of applying deep learning to biological data within Galaxy, helping users understand both the workflow and how to interpret model results responsibly.

[shiltemann]

Thanks @paulocilasjr!

[paulocilasjr]

@shiltemann Thank you so much for the inputs/comments. They were spot on.

[shiltemann]

Thanks a lot for the changes @paulocilasjr, looks good :)

From a technical point of view the FAQs are the only thing left to do (the ones in faqs/index.md will be ignored, so either simply remove them or move each to its own .md file in the faqs folder)

I will let @anuprulez (or other maintainers) review the scientific content