arXiv presents biology-informed framework for evaluating synthetic ovarian cancer multi-omics datasets
Original title: Synthesis of Ovarian Cancer Multi-Omics Datasets: A Biology-Informed Evaluation Framework
The framework evaluates paired RNA expression and CNV profiles from the DECIDER cohort using variational autoencoders.
Paired multi-omic cancer cohorts are costly to assemble and difficult to share. The authors introduce a biology-informed evaluation framework for synthetic data generation based on variational autoencoders. The model is applied to the longitudinal DECIDER high-grade serous ovarian cancer cohort and shows strong fidelity and coherence measures.
- Framework assesses modality-specific fidelity using held-out data agreement.
- Three multi-omic integration architectures are adapted based on variational autoencoders.
- Strongest model shows best training-data proximity.
From the original article
1 min read at arxiv.orgPaired multi-omic cancer cohorts are costly to assemble and difficult to share, motivating synthetic cohorts as a complementary resource. As generative model development accelerates, we shift the focus from proposing yet another architecture to defining evaluation criteria that determine whether synthetic data preserve biologically relevant structure.
By Déborah Boyenval · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.