Research map / Probabilistic and scientific modeling / Molecules and proteins
G2PDiffusion: Cross-Species Genotype-to-Phenotype Prediction via Evolutionary Diffusion
Mengdi Liu, Zhangyang Gao, Hong Chang, Stan Z. Li, Shiguang Shan, Xilin Chen
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 environment-aware MSA conditioner (vs DNABERT) (ablation) | G2PDiffusion baseline · BIOSCAN-5M seen, top-1 | Success rate: 11.50% -> 16.73% | Improved |
| add environment-aware MSA conditioner (vs DNABERT) (ablation) | G2PDiffusion baseline · BIOSCAN-5M seen, top-1 | CLIBDScore: 0.100 -> 0.125 | Improved |
| add dynamic alignment sampling (ablation) | G2PDiffusion baseline · BIOSCAN-5M seen, top-1 | Success rate: 11.50% -> 27.14% | Improved |
| add dynamic alignment sampling (ablation) | G2PDiffusion baseline · BIOSCAN-5M seen, top-5 | PES: 0.187 -> 0.268 | Improved |
| add both conditioner and alignment (ablation) | G2PDiffusion baseline · BIOSCAN-5M seen, top-5 | Success rate: 26.60% -> 65.80% | Improved |
| add both conditioner and alignment (ablation) | G2PDiffusion baseline · BIOSCAN-5M seen, top-1 | CLIBDScore: 0.100 -> 0.182 | Improved |
| retrieve 1 MSA sequence instead of none (m 0 -> 1) (ablation) | G2PDiffusion · BIOSCAN-5M seen, top-1 | Success rate: 27.17% -> 36.23% | Improved |
| retrieve 1 MSA sequence instead of none (m 0 -> 1) (ablation) | G2PDiffusion · BIOSCAN-5M seen, top-5 | Success rate: 53.10% -> 65.80% | Improved |
| increase retrieved MSA size from 2 to 3 (ablation) | G2PDiffusion · BIOSCAN-5M seen, top-1 | Success rate: 31.70% -> 29.90% | Hurt |
| increase retrieved MSA size from 2 to 3 (ablation) | G2PDiffusion · BIOSCAN-5M seen, top-5 | PES: 0.291 -> 0.271 | Hurt |
| use MSA (m=1) instead of none (ablation) | G2PDiffusion · BIOSCAN-5M seen, top-1 | PES: 0.151 -> 0.143 | Mixed |
| G2PDiffusion vs best baseline ControlNet | G2PDiffusion · BIOSCAN-5M seen, top-5 | Success rate: 39.1% -> 65.8% | Improved |
Closest papers by meaning
- Deep Generative Models for Phylogenetic Inference with Complex Evolutionary Processes (NeurIPS 2026 · similarity 0.84)
- BioCLIP 2: Emergent Properties from Scaling Hierarchical Contrastive Learning (NeurIPS 2025 · similarity 0.83)
- What Do You See in Common? Learning Hierarchical Prototypes over Tree-of-Life to Discover Evolutionary Traits (ICLR 2025 · similarity 0.83)
- Towards the Efficient Inference by Incorporating Automated Computational Phenotypes under Covariate Shift (ICML 2025 · similarity 0.83)
- PhyloGen: Language Model-Enhanced Phylogenetic Inference via Graph Structure Generation (NeurIPS 2024 · similarity 0.83)
- BioTrove: A Large Curated Image Dataset Enabling AI for Biodiversity (NeurIPS 2024 · similarity 0.82)
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.
