Quantum Machine Learning for Multi-omics Analysis (ISMB 2025)#

_static/qml_multiomics.png

ISMB/ECCB 2025 Tutorial

Hands-on introduction to quantum machine learning (QML) for multi-omics data using QBioCode.

Overview#

This tutorial introduced:

  • Foundations of quantum computing and QML

  • Benchmarking quantum vs classical ML models

  • Understanding data complexity in biological datasets

  • Practical pipelines using QBioCode tools

Materials#

Agenda#

Format

Expand each session to access description, notebooks, and background material.

Session I — Foundations#

1. Introduction to Quantum Computing

Overview of quantum computing concepts and applications in computational biology.

Materials - 📄 Slides: TBD - 📘 Background: TBD

2. Quantum Machine Learning Basics

Introduction to QML frameworks and their application to classification problems.

Materials - 📄 Slides:

Session II — Data & Benchmarking#

1. Artificial Data Generation

Generate synthetic datasets with controlled complexity for benchmarking ML models.

Materials - 💻 Notebook: https://ibm.github.io/QBioCode/tutorials/Artificial_data_generation/example_data_generation.html

2. QProfiler — Benchmarking ML Models

Automated benchmarking of classical and quantum models: - evaluation across datasets - comparison of embeddings and algorithms

Materials - 💻 Notebook: Tutorial QProfiler - 📄 Slides: QBC - 📘 Documentation:

Session III — Model Selection & Meta-learning#

1. QSage — Meta-learning for Model Selection

Predict which ML models perform best using data complexity features.

Materials - 💻 Notebook: Tutorial QSage - 📄 Slides: QBC - 📘 Documentation: QSage

2. Interpreting Model Performance

Analyze benchmarking results to understand:

  • relationships between dataset complexity and model performance

  • when quantum models outperform classical ones

Materials - 📄 Slides: Characterizing Data Complexity in Machine Learning - 📘 Background: Data complexity documentation

Session IV — Quantum-enhanced Pipelines#

1. Quantum Projection Learning (QPL)

Apply quantum feature maps to enhance classical ML performance.

Materials - 💻 Tutorial: https://ibm.github.io/QBioCode/tutorials/Quantum_Projection_Learning/QPL_example.html

2. Hybrid Quantum-Classical Workflows

Build complete pipelines combining quantum and classical models.

Session V — Wrap-up & Discussion#

1. Discussion and Q&A

Open discussion on limitations, best practices, and future directions.

2. Concluding Remarks

Summary and outlook on quantum computing in life sciences.

Audience#

  • Computational biologists

  • Bioinformaticians

  • Data scientists

Interested in:

  • Quantum computing

  • Machine learning for omics data

Reproducibility#

Run all materials locally:

git clone https://github.com/IBM/QBioCode.git
cd QBioCode
pip install -e .