A central-dogma-guided Transformer-MoE with directed attention and masked reconstruction improves cross-cohort multi-omics performance in cancer subtyping, metastasis detection, and survival prediction.
For each upstream driver node, we perturb the corresponding omics feature and rank downstream RNA or protein targets by the magnitude of their reconstructed response
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
DoGMA: A Central-Dogma-Guided Foundation Model for Multi-Omics Alignment and Multi-Task Learning in Oncology
A central-dogma-guided Transformer-MoE with directed attention and masked reconstruction improves cross-cohort multi-omics performance in cancer subtyping, metastasis detection, and survival prediction.