Research map / Probabilistic and scientific modeling / Molecules and proteins
Towards Universal Gene Regulatory Network Inference: Unlocking Generalizable Regulatory Knowledge in Single-cell Foundation Models
Jiaxin Qi, Hang Li, Yan Cui, Yuhua Zheng, Jianqiang Huang
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 |
|---|---|---|---|
| map frozen scFM gene embeddings to links (Baseline-Emb) instead of attention (Origin-Attn) | scBenchmark · UGRN trained on hESC (Str), scBenchmark | AUPRC: 0.630 -> 0.663 | Improved |
| add virtual value perturbation (VVP) over zero-out perturbation | Baseline-Pert · UGRN trained on hESC (Str), scBenchmark | AUPRC: 0.596 -> 0.663 | Improved |
| add gradient trajectory (GDT) over VVP | VVP · UGRN trained on hESC (Str), scBenchmark | AUPRC: 0.663 -> 0.738 | Improved |
| ensemble VVP, GDT and Emb features (Ens) over GDT alone (ablation) | GDT · UGRN trained on hESC (Str), scBenchmark | AUPRC: 0.738 -> 0.768 | Improved |
| use GDT instead of Baseline-Emb | Baseline-Emb · Spc network generalization | AUPRC: 0.540 -> 0.702 | Improved |
| ensemble features (Ens) over GDT alone on Spc network (ablation) | GDT · Cross-network, Spc | AUPRC: 0.702 -> 0.672 | Hurt |
| use Ens instead of Baseline-Emb when training on L/G mESC | Baseline-Emb · L/G mESC training, scBenchmark | AUPRC: 0.515 -> 0.668 | Improved |
| use scGPT instead of scBenchmark as frozen scFM for Ens (ablation) | Ens · UGRN trained on hESC (Str), scBenchmark | AUPRC: 0.768 -> 0.743 | Hurt |
| raise negative-to-positive ratio from 1 to 10 (ablation) | Ens · Class imbalance, scBenchmark | AUPRC: 0.768 -> 0.413 | Hurt |
| raise negative-to-positive ratio from 1 to 10 (ablation) | Ens · Class imbalance, scBenchmark | AUROC: 0.704 -> 0.706 | No clear effect |
| use Ens instead of Baseline-Emb at N/P=10 | Baseline-Emb · Class imbalance N/P=10 | AUPRC: 0.264 -> 0.413 | Improved |
| train on different species and test on the other (mouse-trained, tested on human) | Ens · Cross-species, scBenchmark | AUPRC: 0.704 -> 0.936 | Improved |
Closest papers by meaning
- KINDLE: Knowledge-Guided Distillation for Prior-Free Gene Regulatory Network Inference (NeurIPS 2025 · similarity 0.91)
- InfoSEM: A Deep Generative Model with Informative Priors for Gene Regulatory Network Inference (ICML 2025 · similarity 0.88)
- Gene Regulatory Network Inference in the Presence of Selection Bias and Latent Confounders (NeurIPS 2025 · similarity 0.88)
- Pruning neural network models for gene regulatory dynamics using data and domain knowledge (NeurIPS 2024 · similarity 0.87)
- Gene Regulatory Network Inference in the Presence of Dropouts: a Causal View (ICLR 2024 · similarity 0.86)
- A Generic Family of Graphical Models: Diversity, Efficiency, and Heterogeneity (ICML 2025 · similarity 0.86)
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.
