Multi-Omics Deep Archetypal Analysis
Finding a way to keep interpretability in highly non-linear dimensionality reduction
High-throughput multi-omic molecular profiling allows probing biological systems at unprecedented resolution. However, the integration and interpretation of high-dimensional, sparse, and noisy multimodal datasets remains challenging. Deriving new biology using current methods is particularly difficult because they are not based on biological principles, but instead focus exclusively on a dimensionality reduction task. We have developed introduce MIDAA (Multiomic Integration with Deep Archetypal Analysis), a framework that combines archetypal analysis, an approach grounded in biological principles, with deep learning. Using the concept of archetypes that are based on evolutionary trade-offs and Pareto optimality – MIDAA finds extreme data points that define the geometry of the latent space, preserving the complexity of biological interactions while retaining an interpretable output.
