Research map / Probabilistic and scientific modeling / Molecules and proteins
CDBridge: A Cross-omics Post-training Bridge Strategy for Context-aware Biological Modeling
Chang Yu, Siyuan Li, Zicheng Liu, Jingbo Zhou, Xianglong Guo, Kai Yu, Yuqing Zhou, Ken Li, Zelin Zang, Zhen Lei, Stan Z. Li
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add ToMe adaptive token merging (ablation) | CDBridge stage 1 · GTEx segmentation | AUC: 0.848 -> 0.882 | Improved |
| add ToMe adaptive token merging (ablation) | CDBridge stage 1 · GTEx expression | R2: 0.021 -> 0.205 | Improved |
| add fixed protein cluster on top of ToMe (ablation) | CDBridge stage 1 · GTEx segmentation | AUC: 0.882 -> 0.990 | Improved |
| make protein cluster learned instead of fixed (ablation) | CDBridge stage 1 · GTEx segmentation | F1: 0.602 -> 0.635 | Improved |
| make protein cluster learned instead of fixed (ablation) | CDBridge stage 1 · GTEx segmentation | AUC: 0.990 -> 0.993 | No clear effect |
| add tissue conditioning (ablation) | CDBridge stage 2 · GTEx expression | R2: 0.215 -> 0.387 | Improved |
| add tissue conditioning (ablation) | CDBridge stage 2 · GTEx expression | Spearman: 0.483 -> 0.618 | Improved |
| tissue embedding alone without DNA features (ablation) | CDBridge control · GTEx expression | R2: 0.387 -> 0.020 | Hurt |
| use CDBridge vs Evo2 backbone alone | DNA encoder · GTEx expression average | R2: 0.021 -> 0.387 | Improved |
| use CDBridge vs AlphaGenome | expression prediction · GTEx expression average | Spearman: 0.438 -> 0.618 | Improved |
| use CDBridge vs Isoformer official (TSS aligned) | expression prediction · GTEx expression average | R2: 0.530 -> 0.387 | Hurt |
| use CDBridge vs LucaOne | isoform retrieval · Isoform Retrieval | MRR: 0.354 -> 0.436 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DNABERT2 — DNABERT-2: Efficient Foundation Model and Benchmark For Multi-Species Genomes (ICLR 2024) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- What DNA Foundation Models Learn Beyond Sequence Composition (NeurIPS 2026 · similarity 0.85)
- GeneZip: Region-Aware Compression for Long Context DNA Modeling (NeurIPS 2026 · similarity 0.85)
- Context-Aware Regularization with Markovian Integration for Attention-Based Nucleotide Analysis (NeurIPS 2025 · similarity 0.85)
- A Cross Modal Knowledge Distillation & Data Augmentation Recipe for Improving Transcriptomics Representations through Morphological Features (ICML 2025 · similarity 0.85)
- Multi-modal Transfer Learning between Biological Foundation Models (NeurIPS 2024 · similarity 0.85)
- A Simple and Comprehensive Benchmark for Single-Cell Transcriptomics (AAAI 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
